International Journal of Disaster Risk Reduction 31 (2018) 806–831

Contents lists available at ScienceDirect

International Journal of Disaster Risk Reduction

journal homepage: www.elsevier.com/locate/ijdrr

Risk, Reliability, Resilience (R3) and beyond in dam engineering: T

A state-of-the-art review
⁎
Mohammad Amin Hariri-Ardebili
Post-Doctoral Research Associate and Adjunct Lecturer, Department of Civil Environmental and Architectural Engineering, University of Colorado, Boulder, USA

A R T I C LE I N FO A B S T R A C T

Keywords: Dams are critical infra-structures whose their failure could leads to high economic and social consequences. For
Dams this reason, application of quantitative risk analysis has gained extensive attention in recent years. Dam safety
Sustainability management has become an indispensable part of all dam engineering projects worldwide. The concept of risk is
Fragility heavily tied to probabilistic methods. From an engineering point of view, a clear definition of the terminologies
Probabilistic
involved in dam safety, and a comprehensive state-of-the-art review of the current literature are the starting
Failure
Safety
points towards an effective risk-based approach. The first part of this paper provides a systematic review on the
Hazard fundamental elements in uncertainty quantification. Then, different terminologies in risk-based dam safety are
explored and their inter-connections are discussed. More than 350 papers are summarized, and several tables
and conceptual plots are used for extra clarification. Since no such a paper is ever published, hopefully this can
unify all the future activities and improves our understanding from probabilistic risk analysis.

1. Introduction high-hazard potential dams. It is not surprising that the Association of

Although dams provide significant benefits to the society, their
impacts on the surrounding environment cannot be neglected, which
include resettlement, socio-economic impacts, environmental concerns,
sedimentation issues, and safety aspects [1]. These concerns should be
reduced through a set of inter-organizational planning and the in-
corporation of different mitigation techniques. Moreover, the failure of
dams, although rare, can be catastrophic and may cause immense da-
mage and loss of life. Therefore, the society demands an increase in
safety of hydropower systems. The most effective way to address these
expectations is to integrate dam's design, construction and operation in
the framework of systematically risk management including the aspects
such as sustainability, resilience, and public participation [2].
Hydropower is the biggest source of the renewable energy in the US,
as of 2017, and over 90,580 dams operate across the country. Based on
the Federal Emergency Management Agency (FEMA) classification, the
overall number of high-hazard dams was about 15,500 as of 2016,
while those considered deficient high-hazard has climbed to about 2200
[3]. Fig. 1 shows a map of all the country's dams. The average age of
dams in the US is over 56 years old according to the American Society
of Civil Engineers (ACSE) 2017 Report Card for America's Infrastructure.
Dams as a whole got a “D” grade in the report card. By 2025, 70% of
dams in the US will be over 50 years old. It is estimated that it will
require an investment of nearly $45 billion to repair aging, yet critical,
State Dam Safety Officials (ASDSO) estimates that it will require a
combined total investment of $64 billion to rehabilitate the nation's
non-federal and federal dams. The US Army Corps of Engineers
(USACE) estimates that more than $25 billion will be required to ad-
dress Corps-owned dam deficiencies, while the US Bureau of
Reclamation (USBR) will need to invest $2 billion over the next 15
years to upgrade 20 of its dams. Given the limited budget for repair and
maintenance, national codes require a comprehensive emergency ac-
tion plan for assessment of dam safety [5–7], and [8].
To this end, investigation of the current and future risks is an im-
portant task in dam safety management. The concept of risk is heavily
tied to probabilistic methods. Developing any comprehensive frame-
work for dam safety management requires a strong theoretical back-
ground in conjunction with appropriate field measurements. From an
engineering point of view, a clear definition of the terminology of the
probabilistic concepts and a comprehensive state-of-the-art review of
the current literature are the starting points to develop a big picture in
dam safety probabilistic assessment.
The probabilistic approach includes but is not limited to un-
certainties in demand, capacity, analysis type, interpretation of the
results, etc. The concepts involved in a probabilistic analysis are sche-
matically shown in Fig. 2, and will be explained later. This paper re-
views the mathematical aspects in uncertainty quantification and ter-
minologies in risk-based dam safety evaluation. Detailed dam potential

⁎
Corresponding author.

https://doi.org/10.1016/j.ijdrr.2018.07.024
Received 21 May 2018; Received in revised form 25 July 2018; Accepted 26 July 2018
Available online 01 August 2018
2212-4209/ © 2018 Elsevier Ltd. All rights reserved.
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

method [20], and finite element method [21].

Usually an smart commercial (e.g. [22,23]) or in-house (e.g.
[24,25]) software is needed in order to develop the finite element
model of the dam and its surrendering environment. The dynamic in-
teraction between the dam, the reservoir water, and the foundation
rock was of primary concern in the coupled system mechanics. Apart
from the Westergaard [26], Zangar [27] approximate solution for the
fluid-structure interaction (FSI), Chopra was the first researcher to
systematically study this problem and provide some simplified for-
mulas, found in [28,29]. Chopra and his colleagues expanded these
studies to include the soil-structure interaction (SSI) effects as well
[30–32]. Nowadays, the FSI and SSI problems are well-established, and
Fig. 1. A map of all US dams color-coded by how hazardous they are; adopted
there have been proposed many advanced techniques to properly and
from FEMA [4].
accurately account for their effects on dams, such as [33–41]. For the
past decades, nonlinear response of dam subjected to applied (external
or internal) stressors has been another hot topic. Nonlinear responses
Robustness are mainly originated from mass concrete cracking (e.g. [42–47]) and/
Uncertainty Sustainability or discrete joints/interface (e.g. [48–52]). Reinforcement is also mod-
Loss
eled in some cases [53–56]. A very detailed and comprehensive
Damage guideline for nonlinear dynamic finite element analysis of concrete
Risk Safety dams can be found in [57].
Reliability All these nonlinearities, along with FSI and SSI, should be combined
Randomness with an appropriate analysis technique to be implemented in a dam
Epistemic Aleatory
Probabilistic engineering problem. FEMA developed a general guideline for Potential
Analysis Failure Mode Analysis (PFMA) of concrete dams, see Table 1, in which
Failure
Modes the level of applicability and reality of each method is illustrated [58].
Vulnerability
Hazard
Resilience Traditionally, evaluation of structural adequacy has been expressed
C
by a factor of safety FS = D , where C is the capacity (i.e., strength) and
Fragility D is the demand (i.e., load) [59]. Although this evaluation is quite
Failure Data simple to understand, it suffers from many limitations: it 1) treats all
Monte Carlo Probability Sampling
Simulation loads equally, 2) does not differentiate between capacity and demands
respective uncertainties, 3) is restricted to service loads, and 4) does not
allow comparison of relative reliabilities among different structures for
Fig. 2. Inter-connection of concepts incorporated in probabilistic analysis. different performance modes. Another major deficiency is that all
parameters are assigned a single value in an analysis which is then
failure modes [9], failure process analysis [10], dam breaching flood deterministic. Based on Ruggeri [60], the Limit Equilibrium Method
[11,12], emergency management, etc. are not discussed in this paper. (LEM) is the most well-consolidated technique for deterministic safety
assessment of concrete dams in which the dam is assumed to be a rigid
2. Transition from deterministic to probabilistic body [61,62]. The main disadvantage of this method is that it is not
calibrated against a useful safety level [63]. Pourkamali-Anaraki [64]
A majority of the research on dam engineering in the past half a formulated the LEM in the following form:
century was focused on the numerical simulation of the dam structure.
FS = f (T , W , U , φ , c, A, α, FSφ, FSc ) (1)
Different techniques were used in these early developments, such as the
finite difference method [13] and the finite element method [14], and where T is shear force; W is the weight; U is uplift force; φ and c are
were gradually expanded to the newer techniques, such as the discrete angle of friction and cohesion at the considered plane, respectively; A is
element method [15], particle finite element method [16], mesh-free the area of rapture; α is the inclination of the sliding with respect to
method [17], and extended finite element method [18,19]; or even horizon; and finally, FSφ ≥ 1 and FSc ≥ 1 are the partial factor of
more mature version of the older techniques such as finite difference safeties with respect to friction and cohesion, respectively. Special

Table 1
Limitations of structural analysis methods to compute the dam response; adopted from FEMA-PFM [58].
Method Can upstream stress be Can the analysis show if Can analysis show if the Can analysis show how far Can the analysis perform a
computed? the dam cracks through? dam slides given loads and the dam can slide? post-seismic stability study?
shear strengths?

Limit Equilibrium Yes, assumes plane sections Yes, pre-supposes crack Yes, it can determine if Yes, using a Newmark type Yes, only sliding factor of
remain plane, so lacks accuracy sliding commences approach safety computed
2D FEM Linear Yes, includes flexibility but no 3D No, computed stresses No, only sliding factor of No, unrealistic linear No, only sliding factor of
effects can be higher than the safety computed by strength carries all imposed safety computed by
strength integrating stresses loads integrating stresses
2D FEM Nonlinear Yes, includes 2D flexibility and Yes, but 3D effects not Yes, but 3D effects not Yes, but 3D effects not Yes, but 3D effects not
nonlinear but no 3D effects included included included included
3D FEM Linear Yes, includes 3D flexibility but no No, computed stresses No, only sliding factor of No, unrealistic linear No, only sliding factor of
nonlinearity can be higher than the safety computed by strength carries all imposed safety computed by
strength integrating stresses loads integrating stresses
3D FEM Nonlinear Yes, best available, includes 3D Yes, best available Yes, best available Yes, best available Yes, best available
flexibility and nonlinearity

807
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

attention should be paid to the fact that a larger FS does not necessarily The swing for each of the n RVs is computed as
imply a smaller risk [65], because its effect can be negated by the
presence of larger uncertainties in the design environment. Yudelevich Θiswing = Θimax − Θimin (5)
[66] discussed on the impact of initial uncertain information on the and are sorted in descending order. Finally, the Tornado diagram is
reliability assessment of concrete gravity dams. He showed that through plotted, and one has to arbitrarily decide what are the most sensitive
a transition from a deterministic to probabilistic method, and further on RVs, Fig. 4. Application of SA and Tornado diagram on dams can be
to a possibility method a correct reliability assessment can be achieved found in [71] for linear and [72] for nonlinear models.
under the conditions of high levels of uncertainty in the initial data.

3.3. Uncertainty analysis

3. Elements of uncertainty quantification
The input factors of models are unfortunately not always known
3.1. Classification of random variables
with a sufficient degree of certainty, which can be caused by natural
variations, as well as by errors and uncertainties associated with mea-
The random variables (RVs) can be categorized as [67]:
surements. The uncertainty of the input parameters is often expressed
Epistemic uncertainty is the scientific uncertainty in the model of
in terms of probability distributions; it can also be specified by samples
the process and is due to limited data and knowledge. For continuous
of measured values, i.e. empirical probability distributions. The un-
random variables, the epistemic uncertainty is modeled by alternative
certainties of the different input parameters may have dependencies on
probability density functions (PDFs) and is usually addressed through
each other, i.e. they may be correlated. Generally, the main reason to
use of a logic tree.
perform an uncertainty analysis is to assess the uncertainty in the model
Aleatory variability is the natural randomness in a process. For
output that arises from uncertainty in the input RVs.
continuous variables, the randomness is parameterized by the PDF.
Lexical uncertainty is due to subjective definition of parameters,
3.4. Sampling techniques
hence, shows the lack of definitive or sharp distinction [68].
The basic qualifier refers to directly observable quantities, such as
Sampling of the distributional model is indeed a key element of an
material properties (e.g., strength, stiffness), loads (e.g., earthquake
magnitude, sea wave height), environmental phenomenons (e.g., tem- uncertainty quantification (UQ). Some of the fundamental techniques
are:
perature, alkali-aggregate reaction), and geometric dimensions (e.g.,
section size). Fig. 3 compares the epistemic and aleatory uncertainties
applied to the seismic hazard analysis. Note that a detailed discussion • Monte Carlo Simulation (MCS): MCS is the most widely used sam-
on the epistemic and aleatory uncertainties in risk analysis of dam pling method. Success of the crude MCS depends on a very large
safety can be found in Baecher [69]. number of analyses. As a palliative to this handicap, an improved
sampling method is achieved through the so-called Latin Hypercube
Sampling (LHS) method [73].
•
3.2. Sensitivity analysis
Pseudo-random method: LHS guarantees samples to be drawn over
the whole range of the distribution. First, the range of each variable
Sensitivity analysis (SA) determines the impact of a variation in an
is split into m non-overlapping intervals of equal marginal prob-
input parameter on output results. Mathematically, this corresponds to
ability 1/ m . Then, sampling starts with the random selection of an
the partial derivative of the output function with respect to an input
interval, followed by another random selection of a point inside it.
•
parameter at a given design point. The procedure starts with the
Quasi-random method: Quasi-random sequences are those gener-
identification of the basic RVs, X = (X1, …, Xn ) , and their corresponding
ated from a completely deterministic low-discrepancy process and
distributional model (e.g., normal and log-normal). Then, 2n + 1 ana-
have no inherent statistical properties. Two famous members in this
lyses are performed [70] using mean ( Ximean ), minimum ( Ximin ), and
group are Halton [74] and Sobol [75]. Fig. 5 compares crude MCS
maximum ( Ximax ) values of the RVs. The response can be mathemati-
vs. quasi-random methods, where the two RVs X1 and X2 have a
cally expressed in terms of
uniform model with 400 samples.
Θ = f (X1 , X2 , …, Xi , …, Xn ) (2)

A reference response is first computed in terms of the n RVs in

ΘRef 3.5. Design of experiments
S = {1, 2, …, n} equal to their mean values
Design of experiments (DOE) refers to a statistical technique that
ΘRef = f (Ximean ), ∀ i ∈ S (3) systematically defines the minimum number of experimental sampling
points to optimize the computed or observed responses. There are dif-
Then, 2n analyses are performed, each corresponding to a given
ferent techniques to build a DOE [76]: Randomized Complete Block
maximum or minimum of a RV, while all others are set to their mean
Design, Latin Square Experimental Design, Full Factorial Design, Frac-
value
tional Factorial Design, Central Composite Design, Box-Behnken Ex-
Θimin = f (Ximin , X jmean ), i = RV , ∀ j ∈ S ∧ j ≠ i perimental Design, Plackett-Burman Experimental Design, and Taguchi
Θimax = f (Ximax , X jmean ), i = RV , ∀ j ∈ S ∧ j ≠ i Experimental Design. Some of the widely-used DOE techniques are
(4)
shown in Fig. 6, where the small circles show the location of the

Aleatory Epistemic Fig. 3. Aleatory vs. epistemic uncertainty.

Epistemic =
Annual prob of exceedance

uncertainty bounds

PDF

Aleatory =
Integration over distribution of curve shape Logic tree of technically
expected parameter values defensible interpretations
Acceleration

808
M.A. Hariri-Ardebili
Fig. 4. Sensitivity analysis using Tornado diagram.
reference points.
Outputs from DOE are then combined to form a response surface
meta-model (RSM). Polynomial models are among the widely used
functions for response surface. A quadratic polynomial with cross terms
is usually used for the majority of civil engineering applications [77]:
k k k−1 k
y = β0 + ∑ βi xi + ∑ βii xi2 + ∑ ∑ βij x i x j + ε
i=1 i=1 i=1 j=i+1
where y is the response variable; x i and x j are coded input variables; β0 ,
βi , βii , and βij are the unknown coefficients; ε refers to error term (in the
regression context); and k is the number of factors. The unknown
parameters can be estimated by the least-squares regression method,
which minimizes the sum of squares of the differences between the
actual output and the approximated one.
In a series of papers, Hariri-Ardebili and his co-authors [78–80]
implemented the concept of DOE to address the material and ground
motion uncertainty in arch and gravity dams. In all cases, first a meta-
model is developed using one or more DOE techniques and then, vali-
dated using a large number of finite element simulations.
3.6. Machine learning techniques
Machine learning, a specific type of artificial intelligence that
“learns” as it identifies new patterns in data and create strategies to
improve decision making based on information hidden in huge data
sets. Choosing the right algorithm is a key point in any machine
learning application, and thus understanding the advantages and dis-
advantages of each technique is vital. Three major machine learning
algorithms are:
• Supervised learning where the algorithm produces a function that
maps inputs to outputs. The application can be found in classifica-
tion and regression.
• Unsupervised learning which the input data is not labeled and
does not have a known result. The application can be found in
clustering, dimensionality reduction and association rule learning.
• Semi-supervised learning in which the input data is a mixture of
labeled and unlabeled examples. The application can be found in
classification and regression.
Fig. 7 highlights a number of diverse classes and subclasses of al-
gorithms and approaches applied in machine learning. Many of these
techniques have been already implemented in dam engineering. The
major idea is to monitor the current behavior of dam and predict the
long-term future behavior.
Although machine learning is not the subject of this paper, the most
relevant applications in dam engineering are briefly addressed. Salazar
International Journal of Disaster Risk Reduction 31 (2018) 806–831
et al. [82,83] compared some of the machine learning based predictive
models for dam safety assessment, i.e., random forests, boosted re-
gression trees, neural network, support vector machine (SVM), and
multivariate adaptive regression splines. Saouma et al. [84] used
stepwise linear regression and K-nearest neighbor local polynomial
techniques for prediction of arch dam responses based on pendulum
recordings. Gaspar et al. [85] proposed a probabilistic thermal model to
(6) propagate uncertainties on some RCC's physical properties. Mata et al.
[86] proposed a method based on linear discriminant models for the
construction of decision rules for the early detection of developing
failure scenarios. Su et al. [87] proposed an early-warning model for
dam safety based on SVM method. Dancea et al. [88] proposed a
computer vision technique in order to enhance the visual inspection
process of large concrete dams using a modified fuzzy c-means algo-
rithm. Gu et al. [89] proposed the use of least squares SVM in back
analysis of RCC dams and determination of the complex mechanical
properties. Moreover, Su et al. [90] applied a similar idea for gravity
dams. Cheng and Zheng [91] proposed multi-variate dam safety mon-
itoring models based on latent variables and least-square SVM (to si-
mulate the nonlinear mapping). Rankovic` et al. [92] developed a
support vector regression model in order to forecasting the tangential
displacement of a concrete dam. Su et al. [93] proposed different al-
gorithms to improve the capacity of SVM based monitoring models by
taking into account the time-variant material and load. Yu et al. [94]
analyzed the safety of a RCC dam on the basis of measurements of the
displacements, strains, and stresses in the concrete. A hydrostatic,
temperature, time-displacement model was used to quantify the con-
tributions each element on the dam's displacements. A multiple linear
regression (MLR) was used to train and test the relationships among the
hydrostatic conditions and to evaluate the stability. Cheng et al. [95]
proposed two different back-analysis frameworks based on optimiza-
tion technique and multivariate machine learning models used to de-
termine the material dynamic parameters of concrete gravity dams.
3.7. General papers in uncertainty quantification
Other researchers have also investigated the uncertainty in the
ground motion and/or material models with the application to the
dams; however, they did not specifically derive fragility/vulnerability
curves, nor did they compute the failure probability and associated risk.
In nearly all of the following papers, the IDA technique is used (with
some small variations) to quantify the seismic uncertainty: [96–105].
Leclerc et al. [106] developed a computer program called CADAM
for static and seismic stability evaluation of gravity dams. It is based on
the rigid body equilibrium and beam theory to perform stress analyses,
compute crack lengths, and assess safety factors. Seismic analyses can
be performed using either the pseudo-static or a simplified response
809
M.A. Hariri-Ardebili
Fig. 5. Crude vs. quasi Monte Carlo methods.
spectrum methods. Probabilistic analyses based on MCS are possible as
well.
Wang et al. [107] quantified the bias of dynamic response and the
damage of an arch dam subjected to artificial earthquake excitations.
Gupta and Joshi [108] presented a response spectrum-based stochastic
method for gathering the statistical estimates of various response
quantities of gravity dams. The random vibration theory for computing
the peak factors of ordered peaks in a stationary stochastic process,
wherein the statistical parameters are defined in terms of the first few
moments of the power spectral density function (PSDF) of the process,
is used for this purpose. An analytical model was developed for the
power spectral density function (PSDF) of response quantities in terms
of the modal properties of the dam and the response spectrum-com-
patible PSDFs of input ground acceleration. Liu et al. [109] first pro-
posed a renewed simulation scheme to represent fully non-stationary
seismic accelerations. Then, this technique was then combined with the
probability density evolution method (PDEM) for nonlinear dynamic
reliability analysis of a randomly base-excited concrete gravity dam.
Hariri-Ardebili and Saouma [110] proposed a set of micro, meta and
macro, damage indices for damage assessment of gravity dams. The
proposed damage index was a combination of dissipated energy, crack
length, and critical displacement. Further, Hariri-Ardebili et al. [111]
proposed a large set of damage and performance indices for damage
analysis of arch dams. The endurance time analysis technique was used
for the simulations. Moreover, Ansari and Agarwal [112] proposed
some DIs based on vibration frequency and also factor of safety against
sliding. Dell'Agnese et al. [113] proposed a physical damage index to
assess the check dams based on empirical evidence collected in
Northern Italy. Bybordiani and Arıcı [114] evaluated the use of 2D
modelings for the prediction of the seismic demands on gravity dams.
Different material properties and a large set of ground motions were
used to get the possible bias.
Modeling uncertainties are further considered in the following pa-
pers: [115–119]. In a separate research, Tang et al. [120], Zhong et al.
[121], Yin et al. [122] investigated the concrete heterogeneity (based
on the Weibull distribution law) on the seismic response of dams
modeled with damage plasticity. They reported a non-smoothed stress
distribution in the case that the concrete heterogeneity is modeled.
Furthermore, they compared the cracking pattern; however, no statis-
tical interpretation was drawn. Increasing the heterogeneity index (i.e.
more homogeneous concrete), increases the potential for localized da-
mage. Hariri-Ardebili et al. [123] studied the linear elastic seismic re-
sponse of a gravity dam with random finite element method. Different
correlation length, combination of random variables, 1D vs. 2D random
field, and the impact of dam class are discussed.
Moreover, Malm et al. [124], and Gasch et al. [125] investigated the
crack propagation in a buttress dam due to stochastic variations in
material properties. Probabilistic analyses were performed in both the
local and global models with MCS where each element was randomly
assigned a concrete strength according to an assumed material dis-
tribution. They found that the weak material properties near the crack-
tip are governed the propagation and the trajectory of the crack. They
also showed that the mean crack propagation (inclination and length)
resulted from probabilistic analyses is different from the deterministic
one.
810
International Journal of Disaster Risk Reduction 31 (2018) 806–831
Pereira et al. [126] proposed a probabilistic model of the reservoir
water level of any new dam based only on its geometrical properties,
which provides a low-cost alternative to in-depth hydrological analysis.
4. Risk-based probabilistic framework
Various researchers have focused on an integrated approach for risk
analysis and risk management, [127–130], and [131]. Risk, reliability,
and fragility were within the scope and benchmark workshop by the
International Commission on Large Dams (ICOLD) for several years
(e.g., 2011, 2013, 2015), with the general conclusion that the risk-
based methods are a useful tool to analyze the risk of dam systems,
including downstream consequence assessments and the uncertainty of
structural models [132].
This section considers seven fundamental concepts in risk-based
probabilistic analysis and provides a short but comprehensive descrip-
tion of each one in the context of dam engineering. Each subsection is
followed by a review of some important published papers on that topic.
4.1. Reliability
Structural reliability analysis deals with the quantitative assessment
of the failure probability, Pf , given a model of the uncertainty in the
structural, environmental, and load parameters. The safety margin, Z, is
defined as the difference between resistance, R, and stress, S [133]:
Z (t ) = R (t ) − S (t ) (7)
Failure would occur if Z < 0 . Note that t represents the time-dependent
reliability assessment [134]. Alternatively, R and S can be combined,
and the result is expressed logarithmically as:
R (t )
X (t ) = ln
S (t ) (8)
in which the failure would occur for negative values of X.
Reliability index, β , is used as a relative measure of the reliability or
confidence in the ability of a structure to perform its function in a sa-
tisfactory manner. Probabilistic methods are used to systematically
evaluate uncertainties in parameters that affect structural performance,
and the reliability index can be connected to the risk matrix [135].
In special cases in which both the demand and capacity are RVs and
presented using the normal distributional models, an analytical solution
can be found for reliability index, β . If X is assumed to follow a normal
distribution, then it has a mean value X = ln S
mean
( ) R
and a standard
deviation (STD) σ . Therefore, the reliability index is defined as the
distance between mean performance value and the limit state (LS)
normalized with respect to the standard deviation:
σln2 (R / S )
μln (R / S ) ln μ (R / S ) −
β= = 2
σln (R / S ) 2
⎡ ⎛ σ(R / S ) ⎞ ⎤
ln ⎢1 + ⎜ ⎟⎥
⎢ μ
⎣ ⎝ (R / S ) ⎠ ⎥⎦ (9)
On the other hand, knowing the value of reliability index, the
probability of failure can be computed as:
Pf = Φ (−β ) (10)
where Φ (.) is the standard normal cumulative distribution function
(CDF).
Reliability indices for a number of components, or a number of
modes of performance, may be used to estimate the overall reliability of
a structure/system. Two extreme cases might be considered:
Series System: Such a system will perform poorly if any one com-
ponent perform unsatisfactorily. If a system has n components in series,
the probability of unsatisfactorily performance of the ith component is
pi and its reliability is Ri = 1 − pi , then the reliability of the system, or
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

Two-level full Three-level full Central composite Taguchi design of

factorial design factorial design design experiment

Fig. 6. Design of experiment models.

probability that all the components will perform satisfactorily is
R = R1 × R2 × ⋯×Rn = (1−p1 )(1−p2 )⋯(1−pn )
Parallel Systems: Such a system will only perform unsatisfactorily
if all component perform unsatisfactorily. Hence the reliability is
R = 1 − (p1 ×p2 ×⋯×pn )
Further discussion on failure probability and reliability of series/
parallel dams can be found in Dewals et al. [136], Huaizhi et al. [137].
Reliability indices are a relative measure of the current condition
and provide a qualitative estimate of the structural performance.
(11) Structures with relatively high reliable indices will be expected to
perform well. If the value is too low, then the structure may be classi-
fied as a hazardous one. Numerically, the failure probability can be
estimated using one of the following techniques: direct integration,
(12) first- and second-order reliability methods (FORM/SORM), crude MCS,
LHS, and Importance Sampling (IS).
In the direct integration, the probability of failure in terms of LS

Naive Bayes
Averaged One-Dependence Estimators
Deep Boltzmann Machine
Deep Belief Networks Bayesian Belief Network
Deep Learning Bayesian
Convolutional Neural Network Gaussian Naive Bayes

Stacked Auto-Encoders Multinomial Naive Bayes

Bayesian Network

Random Forest
Gradient Boosting Machines Classification and Regression Tree
Boosting Iterative Dichotomiser 3
C4.5 & C5.0
Bootstrapped Aggregation Ensemble
AdaBoost Decision Tree Chi-Squared Automatic Interaction
Detection
Machine Learning Algorithms

Stacked Generalization
Decision Stump
Gradient Boosted Regression
Conditional Decision Trees
Trees
M5

Perceptron Neural
Back-Propagation Networks Principal Component Analysis
Hopfield Network Partial Least Squares Regression
Sammon Mapping
Multi-dimensional Scaling
Least Absolute Shrinkage and
Selection Operator Projection Pursuit
Dimensionality
Elastic Net Regularization Principal Component Regression
Reduction Partial Least Squares Discriminant
Least Angle Regression
Analysis
Mixture Discriminant Analysis
Cubist
Quadratuc Discriminant Analysis
One Rule
Regularized Discriminant Analysis
Zero Rule Rule System
Flexible Discriminant Analysis
Repeated Incremental Pruning
to Produce Error Reduction Linear Discriminant Analysis

Linear Regression K-Nearest Neighbour

Learning Vector Quantization
Ordinary Least Squares Regression Instance Based
Self-Organizing Map
Stepwise Regression
Locally Weighted Learning
Multivariate Adaptive Regression Regression
Splines K-Means
Locally Estimated Scatterplot K-Medians
Smoothing
Clustering
Expectation Maximization
Logistic Regression Hierarchical Clustring

Fig. 7. A classification of machine learning algorithms; adopted from Brownlee [81].

811
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

function can be expressed as: Different methods with different levels of complexity and accuracy
were proposed for SORM; however, following is a simple but practical
Pf =  [G (X ) ≤ 0] = ∫G≤0 fR (R) fS (S ) dRdS (13) formula on the basis of asymptotic analysis [139]:
N −1
where the randomness of R and S is expressed by probability density
P fSORM = Φ (−βHL) ∏ (1 + βHL κi)−1/2
functions (PDFs) fR and fS . It is possible to determine the failure
i=1 (19)
probability in terms of joint PDF of the all contributing variables as:
where κi are the principal curvatures of the paraboloid. This equation is
Pf = ∫{x : G (x )≤0} fX (x ) dx (14) asymptotically exact (as βHL → ∞ while βHL κi is kept constant) for any
LS function having a single design point. Fig. 8(b) shows the conceptual
Since the integration domain in this equation is only implicitly comparison between FORM and SORM.
available, the direct estimation of the failure probability is very difficult Crude MCS is used to direct calculation of the failure probability
(and usually impossible). based on joint PDF of all the RVs. Since it is based on the theory of large
First order reliability method (FORM) is based on the combination numbers, Nsim → N∞, an unbiased estimator of Pf is given by:
of 1) a local linear approximation of the LS function in a transformed
Nsim
space, and 2) an iterative search for the “design point” [138]. The first 1 Nfail
Pf
MCS
step to define the Hasofer-Lind reliability index, βHL , is to normalize the
=
Nsim
∑ If (xj ) =
Nsim
j=1 (20)
basic RVs, X , by transferring into a standard normal vector, U ,
Fig. 8(a): where Nfail is the number of failed samples, the hat is the sign of esti-
mation, and
Xi − μ Xi
Ui = 1 if G (x ) ≤ 0
σXi (15) If (x ) = ⎧
⎨
⎩ 0 if G (x ) > 0 (21)
where μ Xi and σXi are the mean and STD of the RV Xi . Note that in this
new transferred space μ Zi = 0 and σUi = 1. Finally, the associated generalized MCS based reliability index can
Note that the LS function G (X ) in the original space will take the be expressed as:
form G (U ) in the standard normal space. The Hasofer-Lind reliability
βMCS = −Φ−1 (Pf ) = Φ−1 (1 − Pf )
MCS MCS
index is defined as the shortest distance from the origin in the standard (22)
normal space to the new transferred failure surface G (U ) , (this point is
LHS was developed with the idea to reduce the variance in the crude
called design point, U *):
MCS of the integrand [140]. LHS can be summarized as follows: Given
U * = argmin { u , G (u ) ≤ 0} the range [0, 1] divided into N equally length intervals 1/ N , one point is
u⊂M (16)
randomly selected from each interval forming a sequence of N samples
and subsequently, in H1, {Xi1}, i = 1, 2, …, N . Another sequence can be constructed in-
dependently but the same method as {Xi2 }, i = 1, 2, …, N . These two
βHL = U * = min ( U ) (17) sequences can be paired to form a so-called bi-dimensional space. This
and the “exact” FORM based failure probability for the linear LS method can be expanded until an m-dimensional sequence of N is
functions can be computed as: formed [73,141,142], and [143].
Mathematically, the algorithm can be presented as follows: Let
P fFORM = Φ (−βHL) (18) {πk }, k = 1, 2, …, m be independent random permutations of
Since the FORM linearly approximates the LS function at the design {1, 2, …, N } , each uniformly distributed over all N! possible permuta-
point, it may over- or underestimate the reliability index for the func- tions. Set:
tions with significant curvature. In such a case, the SORM may be used, (πk (i) − 1) + Uik
where the LS function is approximated by a quadratic function at the Xik = , i = 1, 2, …, N ; k = 1, 2, …, m
N (23)
design point (e.g., by a Tailor series expansion).
where Uik are the uniform independent sample points on [0, 1] interval.
The crude MCS needs many samples, as there is no control on the
sampled point. A lot of methods have been proposed to reduce the
number of sampling and, consequently, reduce the variance of the re-
sponses. IS is one of these techniques [144]. Various scholars proposed
different versions and applications of IS such as [145–148], and [149].
The fundamental idea in IS is to concentrate the distribution of
sampling in the most important region. A simple way is to move the
sampling center from the origin in standard normal space to the design
point on the failure function, as shown in Fig. 9. In this method, a new
sampling PDF, hX (x ) (known as sampling density function), is defined
to obtain the samples in the desired region. Thus, the failure probability
is reformulated as:
fX (x )
P fIS = ∫{x : G (x )≤0} hX (x )
hX (x ) dx
(24)
With the similar analogy of Eq. (20), the failure probability can be
approximated as:
Nsim
1 fX (xj )
Pf =
IS
Nsim
∑ If (xj )
hX (xj )
j=1 (25)
An appropriate choice of hX (x ) is key to successful implementation of IS
Fig. 8. Determination of the reliability, FORM vs. SORM. method. Using the FORM results is one method to reach this goal [150].

812
M.A. Hariri-Ardebili
Fig. 9. Comparison of Crude MCS and IS techniques.
Once the fundamental theories in structural reliability are de-
scribed, the rest of this section is dedicated to reviewing the published
research in the field of reliability analysis of dams.
Prendergast [151] presented the first work on determination of the
failure probability in concrete gravity dams. The author provided load
and resistance models to evaluate the safety in terms of the uncertain
parameters (which include the water height and the earthquake ac-
celeration coefficient). Kreuzer and Bury [152] and Bury and Kreuzer
[153] computed the dam's failure probability while the structure was
assumed to be a rigid body under the sliding failure mode. The annual
peak flood and the ground acceleration were modeled based on Gumbel
distribution. Furthermore, Baylosis and Bennett [154] evaluated the
safety of a high gravity dam assuming three failure modes (i.e., sliding,
overturning and over-stressing). On the other hand, Horyna [155] fo-
cused on post-crack dynamic analysis of gravity dams and evaluated the
reliability against sliding using both the analytical and experimental
studies. The ground motion mechanism, peak ground acceleration
(PGA), and reservoir level were assumed to be main RVs.
Code-based reliability assessment is another topic in which
Jeppsson [156] evaluated the safety of a concrete column (as part of an
existing dam) using the reliability method and the Swedish guideline.
the reliability index is formulated as a function of the coefficient of
variation (COV) in the RVs, such as uplift pressure, ice load, and angle
of friction. Kazemi [157] compared the safety of a concrete gravity
block resulting from the allowable stress method and the ultimate limit
state design principle. He found that the reliability method provides a
consistent level of structural performance in the stability analysis of
dams. Saouma [158] combined the reliability index (based on the point
estimate method and Taylor's series) and finite element fracture me-
chanics to compute the safety index in a rehabilitation project. The
initial dam was evaluated based on linear elastic fracture mechanics
(LEFM), while the retrofitted one was based on the nonlinear fracture
mechanics (NLFM).
In a series of papers, Carvajal et al. [159–161] evaluated the relia-
bility of a gravity RCC dam using FORM and MCS. Variability in the
shear parameters were originated from variability in compressive and
tensile strength. Sliding and cracking were considered as LS functions.
Moreover, Peyras et al. [162] developed a combined method of event
tree based risk analysis and reliability techniques for safety analysis of
concrete dams. Papadrakakis et al. [163] and Piliounis and Lagaros
[164] proposed a neural networks based MCS methodology, combined
with nonlinear finite element analysis, for reliability assessment of large
concrete dams. Parameters such as modulus of elasticity, Poisson's ratio,
tensile strength and concrete fracture energy were considered as RVs.
Failure probability and the flood fragility curves were computed as
well. Westberg Wilde and Johansson [165] investigated the reliability-
based safety of concrete dams. In addition, Westberg Wilde and Jo-
hansson [166] computed the system reliability of a spillway with two
monoliths. LSs were sliding at the concrete-rock contact as well as the
rock mass, and adjusted overturning. All the RVs were defined by sto-
chastic distributions. The safety index was calculated based on the
FORM, and the system reliability was approximated by direct
813
International Journal of Disaster Risk Reduction 31 (2018) 806–831
integration of the bivariate normal distribution.
Xu et al. [167] proposed a finite element-based improved reliability
model for estimation of the failure probability in concrete dams. This
technique is validated against the FORM, SORM, weighted regression
method and space reduced weighted regression method. They reported
that the proposed algorithm has a good convergence and stability and
increases the calculation efficiency. Xu et al. [168] developed a method
for functional reliability analysis of concrete gravity dams which uses a
pseudo excitation method to compute the probabilistic characteristics
of the system. Furthermore, Xu et al. [169] improved the capability of
the model in order to determine the failure probability of the interlayer
sliding belts under seismic excitation. This approach is mainly based on
developing a relationship between failure probabilities of specified in-
terlayer sliding belts and its elements using the Markov chain theorem.
Kartal et al. [170–172] proposed a response surface model for re-
liability analysis of concrete-faced rockfill dams. In this model, an ex-
plicit LS function is presented for the system that facilitates the future
large scale MCS. Altarejos-Garcia et al. [173] explained a method to
estimate the conditional probability of dam-water responses using
complex behavior. This techniques was based on the numerical simu-
lation and reliability analysis.
In a collection of papers, Krounis and Johan [174], Krounis et al.
[175,176], Krounis [177] analyzed (both experimentally and numeri-
cally) the sliding stability and failure probability of dams with bonded
concrete-rock interfaces. The heterogeneity of the interface joint was
taken into account, and different spatially correlation lengths were
assumed for strength parameter. Furthermore, they quantified the effect
of correlation between the cohesion and friction angle on the reliability
index.
Chen et al. [178] evaluated the reliability index of an arch dam
subjected to the temperature variation and seismic loads. The dynamic
effects were calculated by the modal superposition response spectrum.
Concrete and rock elastic modulus, as well as tensile and compressive
strength, were assumed to be RVs. Subsequently, a quadratic RSM was
computed for the LS function. In addition, Wang et al. [179] estimated
the probability of dynamical systems' failure by dividing the system
reliability into two phases: failure mode identification, and failure
probability calculation. Earthquake variability, material randomness,
and uncertainty in nonlinear constitutive model (i.e., damage plasticity,
smeared crack, and extended finite element) were taken into account.
Progressive failure analysis at three levels (i.e., the element failure, path
failure, and system failure) was then performed. This technique is si-
milar to that already proposed by Ardebili and Saouma [110] for da-
mage index (DI) calculation in micro, meta, and macro levels of gravity
dams. Fiorentino et al. [180] developed a program for calculation of the
sliding safety factor for gravity dams under static, response spectrum,
and nonlinear dynamic simulations. The program is based on the MCS.
It was developed using an increasing degree of complexity of analysis,
starting from equivalent static analyses to understand the critical
combination of RVs lead to a sliding safety factor less than unit. Then,
those were assigned to a simplified single-degree-of-freedom (SDOF)
system to perform nonlinear dynamic analyses.
Altarejos-Garcia et al. [181] investigated the reliability of the dam-
foundation interface joint failure, including epistemic uncertainty (i.e.
friction angle and cohesion). They compared the conventional uniform
sampling with spatial variability of the strength parameters. Liu and Li
[182] studied time-dependent reliability of gravity dams using gray
theory and a damage mechanics model. Since this method does not
needs iterative simulation, it greatly enhances the computational effort.
Shi et al. [183] studied the reliability of dams with multiple failure
models. In this method, the failure domain is identified by resorting to
lines rather than random points. They compared the results with crude
MCS and showed the efficiency of the line sampling simulation. Su et al.
[184] proposed a non-probabilistic reliability method for safety eva-
luation of arch dams. This technique only requires the intervals of the
uncertain parameters. The concept of reliability is also used by Chiti
M.A. Hariri-Ardebili
et al. [185] for design optimization of gravity dams subjected to
earthquake excitation. They used the subset simulation technique to
optimize the dam shape by minimizing the total cost of concrete for the
given target reliability.
Wang and Ma [186] utilized the design point method in generalized
random space in conjunction with the method of “divided difference” to
calculate the probability of an implicit performance function from a
dam-foundation system. The method can handle the non-normal and
correlated random variables as well. Zhu et al. [187] considered the
randomness in seepage and dam foundation spatial variability, and
presented a FSI stochastic finite element approach. They proposed a
weighted dynamic RSM for reliability analysis. Cordier and Léger [188]
presented a progressive reliability analysis methodology at four com-
plexity levels: 1) deterministic, 2) semi-probabilistic (partial coeffi-
cient), 3) reliability-based adjustable factor of safety (AFS), and 4)
probabilistic. They compared those techniques for the sliding safety
evaluation of a gravity dam and presented results in terms of sliding
factors of safety, allowable water levels, and demand/capacity ratios.
Hariri-Ardebili [189] proposed an analytical failure model for
generic gravity dam classes, which is optimized based on a large
number of nonlinear finite element analyses. A hybrid parametric-
probabilistic-statistical approach was used to estimate the failure
probability as a function of dam size, material distributional models and
external hydrological hazard. Li et al. [190] proposed a penalty func-
tion method for solving nonlinear optimization problems to compute
the reliability indicator and to calibrate the partial factors (avoiding the
complexity associated with the calibration process of the partial factors)
and maintaining the accuracy of the reliability indicator. They applied
this method for sliding stability analysis of a gravity dam. Finally, Jia
et al. [191] proposed a shear criteria for cemented sand, gravel and
rock (CGSR) dams based on the theories of Mohr Coulomb and double-
shear strength failure. LS functions were simply assumed to be uniaxial
compressive and tensile strength. Then, the reliability index was eval-
uated using the first-order second moment combined with the finite
element method.
Another group of researchers combined the concept of fuzzy logic
[192] with structural reliability analysis. Xu et al. [193] proposed an
improved method for fuzzy reliability analysis of deep sliding surfaces
in rock foundation under dams. They discussed the determination of the
PDF of structure and the membership functions for the interval esti-
mation of the probability in the optimized model. Xin and Chongshi
[194] applied the credibility theory into the stability failure analysis of
gravity dam. Stability was evaluated as a hybrid event considering both
fuzziness and randomness of failure criterion, design parameters and
measured data. Furthermore, Cao et al. [195] studied the stability of
high arch dam abutments as a fuzzy random event. The instability risk
ratio models were proposed based on the credibility theory and were
calculated using the MCS and fuzzy random post-processing.
4.2. Hazard
Hazard analysis is the process of recognizing hazards that may arise
from a system or its environment, documenting their unwanted con-
sequences and analyzing their potential causes. There are three objec-
tives for a hazard analysis:
• Hazard identification: To determine the hazards and hazardous
events of the equipment under control and the control system (in all
modes of operation), for all reasonably foreseeable circumstances
including fault conditions and misuse.
• Causes identification: To analyze the event sequences leading to the
hazardous events identified.
• Risk determination: To analyze the risks associated with the ha-
zardous events.
Hazards may be realized or unrealized. A realized hazard has
814
International Journal of Disaster Risk Reduction 31 (2018) 806–831
happened in the past and can therefore be identified from experience.
An unrealized hazard is a potential for a hazardous situation that has
not happened in the past but can be recognized by analyzing the
characteristics of an environment or failure modes of equipment items.
The following is a review of the widely used hazard analysis techniques
[196,197]:
Preliminary hazard analysis: This is a very broad technique and
an initial study that focuses on 1) identifying apparent hazards, 2) as-
sessing the severity of potential mishaps that could occur involving the
hazards, and 3) identifying means for reducing the risks associated with
the hazards. It focuses on identifying weaknesses early on in the life of a
system, thus saving time and money, which might be required for a
major redesign, if the hazards are discovered at a later date.
Preliminary risk analysis: It is a streamlined mishap-based risk
assessment approach. The primary objective of the technique is to
characterize the risks associated with significant loss scenarios. This
team-based approach relies on subject matter experts systematically
examining the issues. The team postulates combinations of mishaps,
most significant contributors to losses and safeguards. The analysis also
characterizes the risk of the mishaps and identifies recommendations
for reducing risk.
What-if/checklist analysis: What-if analysis is a brainstorming
approach that uses loosely structured questioning to 1) postulate po-
tential upsets that may result in mishaps or system performance pro-
blems, and 2) ensure that appropriate safeguards against those pro-
blems are in place. Furthermore, checklist analysis is a systematic
evaluation against pre-established criteria in the form of one or more
checklists.
Failure modes and effects analyses: It is an inductive reasoning
approach that 1) considers how the failure modes of each component
can result in system performance problems, and 2) ensures that ap-
propriate safeguards against such problems are in place. A quantitative
version of this model is known as Failure Modes, Effects and Criticality
Analysis.
Hazard and operability analysis: This technique is an inductive
approach that uses a systematic process for 1) postulating deviations
from design intents for sections of systems, and 2) ensuring that ap-
propriate safeguards are in place to help prevent system performance
problems.
Fault tree analysis: This is a deductive analysis technique that
graphically models how logical relationships between equipment fail-
ures, human errors and external events can combine to cause specific
mishaps of interest, Fig. 10(a). Sample application to the dams and
appurtenant structures can be found in [198,199].
Event tree analysis: This is an inductive analysis technique that
graphically models the possible outcomes of an initiating event capable
of producing a mishap of interest, as shown in Fig. 10(b). Sample ap-
plication to the dams and appurtenant structures can be found in
[200–205].
Relative ranking/risk indexing: It uses attributes of a vessel, shore
facility, port, or waterway to calculate index numbers that are useful for
making relative comparisons of various alternatives.
Coarse risk analysis: It uses operations/evaluations and associated
functions for accomplishing those operations/evolutions to describe the
activities of a type of vessel or shore facility. Then, possible deviations
in carrying out functions are postulated and evaluated to characterize
the risk of possible mishaps.
Pareto analysis: This is a prioritization technique based solely on
historical data that identifies the most significant items among many.
This technique employs the 80–20 rule, which states that of the pro-
blems (effects) are produced by of the causes.
Root cause analysis: Root cause analysis uses one or a combination
of analysis tools to systematically dissect how a mishap occurred (i.e.,
identifying specific equipment failures, human errors, and external
events contributing to loss). Then, the analysis continues to discover the
underlying root causes of the key contributors to the mishap and to
M.A. Hariri-Ardebili
Fig. 10. Logic tree based analysis methods; adopted from Hariri-Ardebili [206].
make recommendations for correcting the root causes.
Change analysis: It systematically looks for possible risk impacts
and appropriate risk management strategies in situations in which
change is occurring.
Common cause failure analysis: This is a specialized approach for
systematically examining sequences of events stemming from the con-
duct of activities and/or operation of physical systems that cause
multiple failures/errors to occur from the same root causes, thus de-
feating multiple layers of protection simultaneously.
Human error analysis: It involves a range of methods from simple
human factors checklists through more systematic analyses of human
actions to more sophisticated human reliability analyses.
Although there are many sources of hazards in dam engineering,
three main ones are: seismic, hydrologic, and aging [134]. This paper is
mainly focused on the seismic hazard.
In structural engineering, seismic hazard refers to an uncertain re-
lationship between some level of seismic intensity measure (IM) and the
frequency or probability of a particular location experiencing at least
that level of excitation. Usually, a probabilistic seismic hazard analysis
(PSHA) is performed to derive the hazard curves. It expresses a plot
where the horizontal axis is the IM at a site and the vertical one is
815
International Journal of Disaster Risk Reduction 31 (2018) 806–831
annual frequency of exceedance, λIM , (inverse of the return period, TR ),
Fig. 11(a). Probability that the ground motion parameter exceeds a
value y from an earthquake of a given magnitude and distance is:
ln y − ln y ⎤
 [Y >y|m , r ] = 1 − Φ ⎡
⎢ σln y ⎥ (26)
⎣ ⎦
Assuming that the earthquake magnitude and distance are the only
RVs (in most of the cases the uncertainty in these two is significantly
larger that the others), the probability that the ground motion para-
meter Y exceeds the particular value y when an earthquake occurs at
the ith seismic source is given by:
 [Y >y] = ∫R ∫M P [Y >y|m, r ] fM (m) fR (r ) dm dr (27)
where f represents the PDF.
One of the PSHA applications is to develop the seismic hazard maps.
To produce the US national seismic hazard maps, hazard curves were
calculated for several 5%-damped spectral acceleration ordinates at
periods of 0.0 (PGA), 0.2, 0.3, 0.5, 0.75, 1.0, and 2.0 s. They are at
annual recurrence rates of N = 0.00211, 0.00103, and 0.00040 per
year, to obtain the corresponding values for 10%, 5%, and 2%
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

Fig. 11. Quantifying the seismic hazard.

probability of exceedance in 50 years (based on Poisson model), Risk = ∫  [Load Events] ×  [Responses|Loads] ×  [Loads, Responses]
Fig. 11(b).
(28)
There are several papers that have specifically studied the seismic
hazard at particular dam sites. Tosun et al. [207] summarized the where  [A|B] is the conditional probability that A is true, given that B is
methods used for the analysis of seismic hazards and total risk and true and  stands for the consequences.
discussed the seismic hazards of 32 large dams constructed on the Eu- The risk concept can be interpreted either as an individual risk
phrates basin (based on the seismic activity of the dam site, the total (which might threaten a dam's ability to deliver its objectives) or the
risk, physical properties, and the location). Similar works can be found overall risk of the system with different dimensions. Fig. 12(a) shows
for deterministic and probabilistic seismic hazard analyses of dam sites the connection between the risk and its dimensions (i.e., performance,
in different countries: for Jordan [208], for USA [209,210], for Mekong time, and cost). Risk management encompasses activities related to
River (China, Thailand, Vietnam,…) [211], for Turkey [212], for Egypt making risk-informed decisions, prioritizing evaluations of risk, prior-
[213], for Spain [214,215], for India [216], for Iran [217], for Italy itizing risk reduction activities, and making program decisions asso-
[218,219], for Greece [220], and for Lebanon [221]. Overall, Wieland ciated with managing a portfolio of facilities. Risk management
[222] discussed the seismic hazard information the dam engineers need
from seismologists and geologists. The main deliveries are response
spectra or uniform hazard spectra and sometimes the acceleration time
histories.
In a series of papers, Wieland [223–227] discussed different features
of the earthquake hazards in large dams, which include ground shaking,
faulting, and mass movements. The concept of the integral dam safety
was explained in terms of the 1) structural safety, 2) dam safety mon-
itoring, 3) operational safety and maintenance, and 4) emergency
planning. Moreover, the seismic design and performance criteria of
dams recommended by the seismic committee of the ICOLD were pre-
sented. Finally, while most of the research on seismic hazard is focused
on large dams, Frigo et al. [228] evaluated the seismic criticality of
small dams based on a multi-criteria analysis with a rapid screening
survey. They claimed that small-sized dams are critical as well (com-
pared to larger ones) because the majority of these constructions are
built close to inhabited zones.
For those readers who are interested in hydrological hazard, a set of
useful and comprehensive information can be found in [229–231], and
[232]. A good review on the flood failure of dams can be found in
[233].

4.3. Risk

In structural safety, risk can be defined as the “measure of the

probability and severity of an adverse effect to life, health, property, or
environment. In the general case, risk is estimated by the combined
impact of all triplets of scenario, probability of occurrence and the as-
sociated consequence.” Quantitative risk assessment (QRA) takes the Fig. 12. Concept of structural safety risk management and its dimensions
following form [234,173]: [235].

816
M.A. Hariri-Ardebili
includes evaluating the environmental, social, cultural, ethical, poli-
tical, and legal considerations of all parts of the decision process,
Fig. 12(b).
Risk assessment is the process of deciding whether existing risks are
tolerable and present risk control measures are adequate, and if not,
whether alternative risk control measures are justified or should be
implemented. Risk assessment incorporates the risk analysis and risk
evaluation phases. Tolerable risk means different things to different
people and organizations. Some focus on economic risks to their com-
panies or organizations (e.g., insurance, offshore oil and gas), while
others focus on loss of life. Most of the regularities use a “risk curve”
[236], either in the form of f-N or F-N chart.
An f-N event chart is composed of individual f-N pairs. Each pair
typically represents one failure mode or the summation of selected
failure modes. f represents the “annualized failure probability” (AFP),
and N represents the expected life loss or number of fatalities. F-N chart
is, in fact, a complementary CDF to portray risk, while the vertical axis
represents the annual exceedance probability of causing N casualties or
greater. Two examples are shown in Fig. 13 from the USBR and Aus-
tralian National Committee of Large Dams (ANCOLD), where “ALARP”
stands for “as low as reasonably practicable”. Bowles [237] defined the
following factors as a metric for judging on whether risks are ALARP: 1)
the level of risk in relation to the tolerable risk limits, 2) the dis-
proportion between the cost of implementing the risk-reduction mea-
sures and the subsequent risk reduction achieved, 3) the cost-effec-
tiveness of the risk-reduction measures, 4) compliance with good
established practice, 5) and societal concerns as revealed by consulta-
tion with the community and other stakeholders.
The UK Health and Safety Executive (HSE) regulates the safety of all
workplace activities in the UK. The HSE was a leader in developing the
general tolerability of risk framework. The HSE proposes that F-N
curves can be helpful for decision-making; however, it does not speci-
fically promote one for water resources structures. Instead, HSE pro-
poses Fig. 14 for individual and societal risks. They are inverted tri-
angles, in which the shape indicates the amount of attention and risk
limits of a particular situation [238]. Except in extraordinary circum-
stances, unacceptable risks must be reduced regardless of the cost of
Fig. 13. Organizational social risk guidelines.
817
International Journal of Disaster Risk Reduction 31 (2018) 806–831
these risk reduction measures. The following limits are recommended
by USACE [239]:
• Annual failure probability (AFP):
- AFP > 0.0001 per year: unacceptable, except in exceptional cir-
cumstances,
- AFP < 0.0001 per year: will be considered tolerable provided other
tolerable risk guidelines are met
• Annual life loss (ALL):
- ALL > 0.01 lives/year: unacceptable except in exceptional cir-
cumstances and is a reason for urgent actions to reduce risk,
- 0.001 < ALL < 0.01 lives/year: unacceptable except in exceptional
circumstances and is a reason for actions taken to reduce risk,
- ALL < 0.001 lives/year: may be considered tolerable provided
other guidelines are met.
Note that criteria for life and economic risk, as well as ALARP, for
China were discussed by Ge et al. [240].
ANCOLD [7] proposes a general framework for risk assessment of
dams as illustrated in Fig. 15. A majority of the research in risk analysis
of dams is based on the hydrologic/flood hazard. Few of them account
for the seismic and aging hazards as well. Additionally, nearly all the
research papers dealing with the risk analysis include a section or
chapter on the calculation of failure probability and probabilistic post-
processing on the data.
Cheng et al. [241] recommended adoption of probabilistic ap-
proaches for safety analysis of US dams. They compared several
methods for estimating dam overtopping risk, including direct in-
tegration, MCS, mean-value first-order second-moment (MFOSM), and
advanced first-order second-moment (AFOSM) analyses. They eval-
uated several geophysical forces which may cause overtopping of a dam
and finally focused on floods and wind, and only touching briefly on
landslides and earthquakes. Fault tree analysis, with the assumption of
Poisson process for extreme events was used to estimate the probability
of overtopping. National Research Council [242] provided a review and
critique of the techniques used to approximate the rare flood prob-
abilities. This information was subsequently used for risk analysis.
M.A. Hariri-Ardebili
Fig. 14. HSE tolerability of risk framework; adopted from HSE [238].
Fig. 15. Typical risk assessment process for a dam; adopted from ANCOLD [7].
Using the QRA, Bowles [243] advocated for employing this approach to
solve the dam safety problems in three levels: 1) determination of the
safety level for low-hazard dams, 2) assessment of the alternative op-
tions for risk mitigation at dams with insufficient spillway capacity, and
3) convincing the decision-makers about the applicability of the usual
design standards.
Stedinger et al. [244] wrote a report explaining the concepts needed
to perform a thorough probabilistic analysis of the dam safety issue,
including the calculation of the probability of dam failure and the
distributions for dollar damages and loss of life. They only considered
the hydrological risk with probable maximum flood (PMF). Bowles
et al. [245] reviewed the relationship between the standard- and risk-
based approaches for dam safety in the context of comprehensive risk
management. They emphasized defining the decision process, the role
of decision criteria, and the involvement of the owners and stakeholder
in a decision-driven and staged-risk assessment process. They also
818
International Journal of Disaster Risk Reduction 31 (2018) 806–831
addressed emergency vs. long term decisions.
Kostov et al. [246,247] computed the quantitative seismic risk of
two concrete gravity dams and one arch dam. A detailed finite element
model was used and the seismic risk was calculated by integrating the
hazard curves with the specific fragility curves, based on LHS model. In
a series of publications, Bowles et al. [248], Bowles [249,250] explains
the step-by-step procedure for portfolio risk assessment (PRA) of dams.
A portfolio is a group of dams for which a single owner or regulator is
responsible for. General steps include: 1) define the overall business
context and objectives for PRA, 2) agree on the business requirements
for PRA, 3) perform an engineering assessment of existing dams, 4)
conduct risk assessment of existing dams, 5) investigate the risk re-
duction alternatives, and 6) integrate dam safety PRA results into
business risk.
Furthermore, Bowles et al. [251] conducted a demonstration risk
assessment (DRA) on a high rolled-earthfill dam. The existing dam and
19 structural risk reduction alternatives were evaluated for flood,
earthquake, and normal operating conditions. The DRA involved an
engineering team and a consequences assessment team. It could provide
an effective way for a dam owner or regulator to gain practical ex-
perience with risk assessment and to evaluate its benefits within the
context of their organizational mission. Rettermeier et al. [252] gen-
erally discusses risk analysis, risk management, and how they are
connected to dam safety in Germany. Richards and Snorteland [253]
had a similar discussion on dam risk management, its challenges, and
current practices in the US. Hill et al. [254] discusses the importance of
the joint failure probability of different dam components. In most cases,
the failure modes of various dam components are not mutually ex-
clusive, and therefore, the failure probabilities are not simply additive.
The total failure probability should be adjusted for these upgrades.
Ignoring adjustments can lead to an overestimation of the total failure
probability and the annual incremental loss of life, and an under-
estimation of the cost per statistical life saved.
Chauhan and Bowles [255] presented a framework for uncertainty
analysis in dam safety risk assessment, including an approach to in-
corporating input uncertainties into the risk analysis model. It provided
some useful formats for representing the uncertainties in risk analysis
results, such as the confidence level associated with meeting tolerable
risk guidelines. Peyras et al. [256] proposed a method to support dam
aging diagnosis and risk analysis that capitalizes on an expert's
knowledge and feedback. The approach can be decomposed into three
parts: 1) an aging functional model based on the failure mode and effect
analysis using a causal graph representation of aging scenarios leading
to loss of functions 2) a qualitative method of describing dam aging
historical data and representing trends in performance loss, and 3)
qualitative methods to assess the risk of performance loss of dams and
their components.
Several software programs were developed to facilitate dam safety
risk analysis. Matalucci [257] described the general framework used in
RAM-DSM (which stands for risk assessment methodology for dams),
developed by Sandia National Laboratories. It is used to analyze the
M.A. Hariri-Ardebili
Fig. 16. General framework of PEER PBEE methodology; adopted from Porter [274].
current security risks of US dams and to provide information to support
effective risk-reduction decisions by dam management. The core of
RAM-DSM is the process by which a systematic risk assessment is ac-
complished. Meghella and Eusebio [258] developed a tool for risk
analysis and prioritization of dams in Italy. In this program, which is
generally based on the event trees scenario, the importance of a dam is
quantified according to failure modes, adverse conditions, defense
groups, physical conditions, and priority indices. Bowles and Aboelata
[259] developed LIFESim, a modular, spatially-distributed, dynamic
simulation system for estimating potential life loss from natural and
dam-failure floods. It provides some effective tools for emergency
planning, including the evacuation time in downstream facilities based
on the deterministic or uncertainty modes. He et al. [260] developed a
dam risk assistant analysis system using Delphi visual programming
language. The program is in the context of the “three-tier and multi-
database” logic structure. Srivastava [261] and Srivastava et al. [262]
developed a computation tool called DAMRAE-U for dam safety risk
assessment. The program is structured to analyze knowledge un-
certainty for the event tree variables and natural variability associated
with flood and earthquake loadings. It also separates the effects of
uncertainty in the existing condition of the dam system by which event
tree model is dependent. i Presas Risk Analysis is a tool for dam risk
calculation, with a graphical user interface designed to integrate
loading probabilities and dam system response using influence dia-
grams [263,264].
Smith [265] proposed a model for dam risk analysis based on
Bayesian networks. It allowed determination of the overall dam risk by
taking into account the numerous inter-relations between the failure
mechanisms, the uncertainties, and the expert judgments. Risk was
characterized without reference to an absolute interpretation of the
dam failure probability, which can cause problems related to
its interpretation and use. They used an interesting version of Bayes’
theorem in the form of:  [Cause | Effect] ×  [Effect] =
 [Effect | Cause] ×  [Cause] . Wu et al. [266] proposed a risk degree
(rate) concept for dams based on monitoring data (recorded from dam
body) and the Bayesian approach. In this technique, the reliability
theory is combined with field monitoring data. A five-grade risk degree
system for dam operation risk and corresponding risk degree was also
defined. Moreover, Wu et al. [267] proposed an operation risk assess-
ment methodology based on the characteristics of major unsafe hy-
droelectric projects. Data was collected from field detection, in situ
monitoring, and regular safety inspection. In addition, a three-layer
hierarchical system was constructed and an improved technique,
combining genetic algorithms and analytical-hierarchical processes was
established.
Serrano-Lombillo et al. [268] proposed a technique to calculate the
incremental risks in the context of an event tree. The main advantage
being that it could be applied to systems of several dams: with a single
risk model that describes the complete system and with a single cal-
culation the incremental risks of the system can be obtained, allocating
the risk of each dam and of each failure mode. Su and Wen [269]
performed a risk analysis combined with fuzzy mathematics to evaluate
819
International Journal of Disaster Risk Reduction 31 (2018) 806–831
the stability of gravity dams. They used membership functions to de-
scribe the extent of stability failure risk, while the fuzziness of both the
design parameters and failure criterion were eliminated through a
“level set” based transformation. Su et al. [270] further studied the
time-varying risk probability of dams based on fuzzy theory.
Cloete et al. [271] presented the rational quantitative optimal ap-
proach, which was a robust risk evaluation model, to produce a defi-
nitive result for the risk mitigation from overtopping of dams. They
provided analyses for a portfolio of dams with discrete optimal results
and not in terms of probability. This process was applied mechan-
istically and did not require judgment from decision-makers. For the
case study, they considered the internal and external risks in embank-
ment dams, but the results are applicable to concrete dams as well.
Brown [272] provided a comprehensive review of the risk mitigation
methods for large concrete dams during design and construction. He
mainly focused on rock engineering features and the potential failure
modes (PFMs). Castillo-Rodríguez et al. [273] proposed a variation of
the combined risk analysis approach for complex dam-levee systems,
which was based on the event tree analysis from multiple combinations
of “load-system response-consequence” events. Serrano-Lombillo et al.
[264] presented the risk reduction indicator “equity weighted adjusted
cost per statistical life saved” (EWACSLS). It allows obtaining prior-
itization sequences of investments while maintaining an equilibrium
between equity and efficiency principles. They applied this technique to
a portfolio of 27 dams where 93 structural and non-structural invest-
ments were prioritized.
4.4. Fragility
The concept of fragility analysis is heavily tied to next-generation
performance-based earthquake engineering (PBEE) [274]. The Pacific
Earthquake Engineering Research (PEER) Center developed this fra-
mework to facilitate direct calculation of the effects of uncertainty and
the randomness of each step in the performance-based procedure,
Fig. 16.
Fragility curve is a continuous function showing the probability of
exceedance of a certain LS for specific level of ground motion IM,
Fig. 17(a) [275]:
Fragility =  [LS|IM = im] (29)
where im refers to a particular value of IM, i.e., there is no uncertainty.
Alternatively, fragility can be defined as [276]:
D
Fragility =  [D ≥ CLS |IM = im] =  ⎡ ≥ 1|IM = im ⎤
⎢ CLS
⎣ ⎥
⎦ (30)
where D is the demand parameter and CLS is the capacity associated
with the given LS. In the context of PBEE, where the structural re-
sponses are expressed as engineering demand parameter (EDP), the
fragility can be defined as:
Fragility =  [EDP≥edp|IM = im] (31)
where edp is a specific threshold value of EDP.
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

interface was modeled by a perfectly plastic Mohr-Coulomb friction

Fig. 17. Fragility function.
Fragility surface is a three-dimensional version of the fragility
curves [277]. Fragility surface can be presented with vectorized IM
(double intensity measures). For such a system the fragility definition
can be replaced by:
Fragility =  [LS|IM1 = im1, IM2 = im2]
The two IMs can be chosen from un-scalable parameters, e.g.
< Mw , Rrup > (earthquake magnitude and distance), or the scalable ones,
e.g. < Sa (T1), IA > (first-mode spectral acceleration and Arias intensity),
Fig. 17(b).
Hariri-Ardebili and Saouma [278] provided a comprehensive state-
of-the-art review on all the published papers in fragility analysis of
concrete dams from 1998 to 2016. The following provides a short
summary on each paper, along with Table 2, which compares and
contrasts their major features.
de Araújo and Awruch [279] [ID-1] is the first published research
addressing the seismic fragility of concrete dams. Both the concrete
properties and the seismic excitation were assumed to be RV and MCS
was used for data sampling. The seismic excitation was artificially
generated using a non-stationary stochastic process. The FS against
sliding, concrete crushing at the toe and concrete cracking (with
smeared crack model) at the heel were computed for 50 simulations. A
Gaussian CDF was fitted to the empirical data point; however, one
should note that the widely accepted form for the fragility is the log-
normal CDF [280]:
ln(x / η) ⎞
 [LS| X= x ] = Φ ⎜⎛ ⎟
⎝ β ⎠
where X is the uncertain excitation. η and β are the median and the
logarithmic standard deviation, respectively. From the method of mo-
ments (MOM), the estimated η and β are [281]:
N
⎛ 1
Nobs
∑i =obs ̂ 2
(ln(Xi) − ln(η ))
η ̂ = exp ⎜
N
∑ ln(Xi) ⎞⎟, β = 1
Nobs − 1
⎝ obs i=1 ⎠ (34)
where Nobs is number of observations.
Tekie and Ellingwood [282] [ID-2] accounted for both the material
and seismic uncertainties along with LHS [283]. Dam-foundation
law. The uplift pressure was modeled by piecewise linear distribution
while the simplified two-parameter Darbre's model was adopted for FSI.
Four LSs were considered as: 1) cracking of the neck, 2) foundation
material compressive failure at the toe, 3) plastic strains of 10 4 and 103 ,
3) sliding at the base, and 4) relative deflection at the crest about 8 and
15 mm.
Lin and Adams [284] and Lin and Adams [285] [ID-3] presented a
set of empirical seismic fragility curves for Canadian concrete dams.
First, the fragility curves were developed for the western side using the
matrices provided by ATC-13 [286]. Then, the curves were modified for
the eastern side using an innovative technique which combines the PGA
and spectral accelerations. The age of the dam was also considered by
introducing standard (built in the 1950 s), old (built in the 1900 s), and
new (built in the 2000 s) construction types.
Mirzahosseinkashani and Ghaemian [287] [ID-4] considered only
the ground motion record-to-record variability for fragility analysis of
PineFlat dam. A concrete smeared crack model was used along with a
linear elastic mass-less foundation. For the seismic analyses, the hor-
izontal component of six near-field ground motions were selected and
each one was scaled from PGA = 0.1–0.7 g. Two LS functions were
considered: 1) crack length at the base, and 2) total cracked area on the
dam's face.
Lupoi and Callari [288,289] [ID-5] used a new procedure in which
the randomness of the external actions was treated separately from the
structural uncertainty. A 46 m high gravity dam was selected, water
modeled by displacement-based fluid elements and the foundation mass
considered along with a bi-linear uplift pressure. Only the linear elastic
analyses were performed for “operational LSs”. Variability in the water
level also was taken into account. Two sets of arrangements were
considered for the dam components: series and parallel. They reported
that series arrangement increases the probability of failure.
(32) Yao et al. [290] [ID-6] only used the record-to-record variability for
fragility analysis of a high arch dam and 72% filled reservoir. Only
three of the contraction joints were modeled, and the mass concrete was
assumed to be linear elastic. A total of 18 ground motions were selected
and used in three groups (each one 2–3 scale levels). Finally, the
maximum sliding and opening at the crest joint were reported without
plotting the actual fragility curves.
Yücel [71] [ID-7] performed a set of parametric linear elastic ana-
lyses on a typical gravity dam. The variables were dam height, down-
stream face slope, ratio of elasticity modulus in concrete to rock, and
concrete tensile strength. Twenty real ground motions were used, and a
total of 2700 analyses were performed. Finally, a Weibull distributional
model was fitted to the data points.
Zhong et al. [54] [ID-8] also considered both the ground motion and
concrete uncertainties using MCS. No details were provided about the
number of ground motions, intensity levels, and type of distributional
models for RVs. A 296 m high arch dam was analyzed and a five-level
damage state (DS) was proposed for data interpretation. Log-normal
CDF was used to fit the data points.
Abdelhamid et al. [291] [ID-9] used a similar technique to Tekie
and Ellingwood [282] to derive the fragility curves of an old gravity
dam subjected to near-fault ground motions. Concrete was linear elastic
and dam-rock interface governed by Coulomb's friction law. LHS was
(33)
used to sample the six uncorrelated RVs. Four LSs were assumed: 1)
tensile stress at the neck, 2) sliding at the base, 3) crest relative dis-
placement, and 4) compressive stress at the heel.
Ju and Jung [292] [ID-10] performed the seismic fragility on a
933 m long (but short) weir overflow with steel reinforcement. The
numerical model was linear elastic and 30 near-field as well as 30 far-
field ground motions were selected and scaled in seven intensity levels.
LSs were defined by the concrete tensile and compressive stresses and
the crest displacement. Finally, a log-normal CDF was fitted to the
empirical data points.
Ghanaat and his co-authors published three papers on the fragility

820
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

Table 2
Summary of the fragility analysis of concrete dams, adopted from Hariri-Ardebili and Saouma [278].
ID Year Dam Software Joint Mat. NL FSI SSI Uplift LE or NL Technique Mat Uncert. Sampling Total Sim. Fragility

1 1998 Gravity Not specified No Yes Yes Yes No NL Random Yes MCS 50 Gussian CDF
2 2003 Gravity Abaqus Yes No Yes Yes Yes NL IDA Yes LHS 72 LN CDF
3 2007–8 All – No No No No No – Empirical No – – LN CDF
4 2009 Gravity NSAG-DRI No Yes Yes Yes No NL IDA No – 42 LN CDF
5 2011–2 Gravity FEAP No No Yes Yes Yes LE Random Yes MCS 60 Empirical
6 2012 Arch Abaqus Yes No Yes Yes No NL MSA-IDA No – 48 –
7 2013 Gravity EAGD-84 No No Yes Yes No LE Single dynamic Yes Parametric 2700 Weibull dist.
8 2013 Arch Not specified No Yes Yes Yes No NL Random Yes MCS Not specified LN CDF
9 2013 Gravity Not specified Yes No Yes Yes No NL Random Yes LHS 60 LN CDF
10 2015 Overflow weir Abaqus No No Yes Yes Yes LE IDA No – 420 LN CDF
11 2011 Turbine blg. LS-Dyna Yes No Yes Yes Yes NL Truncated IDA Yes LHS 150 LN CDF
12 2012 Gravity LS-Dyna Yes No Yes Yes Yes NL Truncated IDA Yes LHS 100 Weibull dist.
13 2015 Overflow Sec. LS-Dyna Yes Yes Yes Yes Yes NL Truncated IDA Yes LHS 300 LN CDF
14 2015 Arch Ansys Yes No Yes Yes No NL IDA No – 80 N CDF
15 2016 Arch Ansys Yes Yes Yes Yes No LE, NL MSA No – 54 LN CDF
16 2016 Gravity Merlin Yes No Yes Yes Yes NL CLA No – 100 LN CDF
17 2016 Gravity Merlin Yes Yes Yes Yes Yes NL Full IDA No – 1200 LN CDF
18 2016 Gravity Merlin Yes No Yes Yes Yes NL ETA Yes LHS 200 LN CDF
19 2015 Gravity LS-Dyna Yes Yes Yes Yes Yes NL Truncated IDA Yes LHS 160 N/LN/Weibull
20 2016 Gravity LS-Dyna Yes Yes Yes Yes Yes NL MSA / IDA Yes LHS 140 LN CDF
21 2016 Gravity Abaqus No Yes Yes Yes Yes NL IDA No – 170 LN CDF
22 – – – – – – – – – – – – – –
23 2016 Gravity – No No No No No LE Random Yes LHS ∞ Empirical
24 2017 Gravity Abaqus Yes Yes Yes No No NL CLA No – 135 LN CDF
25 2018 CFRD GEODYNA No Yes No Yes No NL IDA No – ∼ 100 LN CDF
26 2018 Gravity Abaqus Yes Yes Yes No No NL CLA No – 200 LN CDF
27 2018 Gravity Merlin No No Yes Yes No LE Spectrum Yes LHS, Sobol, 10,000 LN CDF
28 2018 Arch Abaqus Yes Yes Yes Yes No NL IDA Yes LHS 100 LN CDF, Spline
29 2018 Arch Ansys No Yes Yes Yes No NL BBED Yes LHS 275 LN CDF Empirical

Notes: IDA: incremental dynamic analysis; MSA: multiple stripe analysis; CLA: cloud analysis; NL: nonlinear; LE: linear elastic; BBED: Box-Behnken experimental
design; MCS: Monte Carlo Simulation; LHS: Latin Hypercube Sampling; CDF: cumulative distribution function; LN: Log-normal.

curves of different dam types. Their approach is similar to Tekie and
Ellingwood [282]'s, although there are some variations. In all three
papers, both material/modeling and seismic uncertainties were con-
sidered, and LHS was used for data sampling in all cases. In Ghanaat
et al. [293] [ID-11], the seismic performance of a weir structure, in-
cluding a turbine building and an administration building, was studied.
Subsequently, a log-normal CDF was fitted to the data points using the
least-square approach. Ghanaat et al. [294] [ID-12] was again similar to
the first study which was applied on the tallest non-overflow section of
the Folsom Dam. They used the Weibull distribution for curve fitting. In
Ghanaat et al. [295] [ID-13], they derived the fragility curves for a tall
reinforced concrete overflow section of the previous example.
Ghanaat et al. [296] [ID-14] performed fragility analysis on a high
arch dam in which the nonlinearity was originated from the contraction
joints. The incremental dynamic analysis (IDA) technique was per-
formed, while the DI was selected to be a percentage of the overstressed
area on th dam's faces. A 5th order polynomial equation was used to
develop the IDA curves. Three IM parameters were also contrasted:
PGA, peak ground velocity (PGV), and Sa (T1) . They used a normal
distribution for fragility curves and got some non-zero probability of
exceedance at zero IM. Moreover, some of the fragility curves inter-
sected which could be resolved by the recommendation in Porter et al.
[280].
Hariri-Ardebili and his co-authors performed a series of fragility
analyses on gravity and arch dams with different techniques and con-
siderations. Hariri-Ardebili et al. [297] [ID-15] used the multiple stripe
analysis on an arch dam. Only the ground motion variability was con-
sidered for both the linear and nonlinear dam models. In linear analysis,
criteria such as stress, strain, displacement, demand capacity ratio,
cumulative inelastic duration, cumulative inelastic area, and damage
spatial distribution ratio were computed. On the other hand, in the
nonlinear analyses, displacements, joint opening/sliding, and crack-
based DIs were considered. Fragility curves for one stripe were
computed, and a relation was established between the linear and
nonlinear models. Hariri-Ardebili and Saouma [298] [ID-16] performed
the first application of cloud analysis on concrete dams. In this method,
the discrete data points are linear in the logarithmic scale [299,300]:
ln(η (IM)) = b. ln(IM) + ln(a) (35)
where ln(a) and b are the linear regression constants and η is the
median value of EDP given IM.
A two-dimensional finite element model of the non-overflow
monolith of the PineFlat Dam was used, while the only source of non-
linearity being the interface between the dam and foundation. 100
ground motions were selected and the fragility curves and surfaces were
derived using 70 different IM parameters. Hariri-Ardebili and Saouma
[301] [ID-17] developed the collapse fragility curves for gravity dams
based on IDA method. They identified the “resurrection” phenomenon,
as is discussed by Vamvatsikos and Cornell [302]. Two ground motion
combinations were considered: 1) horizontal only, and 2) horizontal
and vertical. Furthermore, two loading scenarios were studied: full and
empty pool. Finally, a log-normal CDF was fitted through the discrete
data points based on three techniques: MOM, sum of squared error
(SSE), and maximum likelihood estimation (MLE) [281].
Hariri-Ardebili and Saouma [72] [ID-18] used the endurance time
analysis (ETA) technique to derive fragility curves. Since a single ac-
celeration function is used, there is no room for record-to-record
variability, and hence, only the epistemic (material/modeling) un-
certainties were considered. A large set of RVs associated with the in-
terface joint, concrete, and rock were taken into account: tangential/
normal stiffness, tensile strength, cohesion, friction/dilatancy angle,
mode I/II fracture energy, modulus of elasticity, and Poisson's ratio.
LHS was used for data sampling with and without correlation among
the RVs. Finally, fragility curves were developed based on log-normal
CDF assumptions and the MLE approach.
Bernier et al. [303] [ID-19] developed the fragility curves based on

821
M.A. Hariri-Ardebili
3D finite element analyses of a gravity dam. Nonlinearity stems from
horizontal joints at the dam-foundation interface and a single lift joint
at the neck. Several synthetic ground motions were generated and ap-
plied. Material uncertainties were assigned only to the interfaces (uni-
form distribution): cohesion, tensile strength, and angle of friction. The
damping ratio was also a RV with log-normal distribution. Three types
of distributional models, e.g. normal, log-normal, and Weibull were
used on the raw data based on two fitting methods: SSE and MLE.
Moreover, Bernier et al. [304] [ID-20] compared the multiple stripe
analysis (MSA) with IDA for the same case study and fitted a log-normal
CDF to the data points. Resulting fragility curves for four LSs with and
without material/modeling uncertainty were contrasted. It was found
that the median and dispersion of the curves are significantly higher
using the MSA method.
Ignoring material/modeling uncertainties, Ansari and Agarwal
[305] [ID-21] accounted only for 17 ground motion variability and
adopted the IDA approach. A concrete damage plasticity model was
used and the weakness of the dam-foundation interface was ignored.
They used four DSs as: slight, moderate, extensive, and severe. Ansari
et al. [306] [ID-22] also derived the IDA-based fragility curves for five
different geometries of concrete gravity dams while the effect of
smoothening of re-entrant corners was discussed. Morales-Torres et al.
[307] [ID-23] discussed a simplified method to build the fragility
curves of sliding failure of concrete gravity dams both natural and
epistemic uncertainties. They considered the limit equilibrium analysis
for the calculations. Yazdani and Alembagheri [308] [ID-24] compared
the pulse-like and non-pulse-like fragility curves for gravity dams using
75 and 60 ground motions respectively. They used the cloud analysis
technique. The rest of the application is identical to that which is re-
ported in [298].
Pang et al. [309] [ID-25] derived IDA-based fragility curves for high
concrete face rockfill dams with a similar technique to that of Hariri-
Ardebili and Saouma [301]. Alembagheri [310] [ID-26] investigated
the efficiency of vector-valued intensity measures for gravity dams and
calculated the fragility curves employing scalar and vector IMs, con-
sidering the effect of zero response values. Permanent deformation and
elemental damage were the damage measures. Hariri-Ardebili and
Saouma [311] [ID-27] proposed a response spectrum-based simplified
analytical model for probabilistic seismic assessment of gravity dams,
including the uncertainty in the material parameters and reservoir
level. They generalized their findings for different dam classes. The
impacts of sampling size and technique (i.e. pseudo- and quasi-random)
were also discussed. Finally, a time-based performance model and fra-
gility curves were derived.
Wang et al. [312] [ID-28] drove the fragility curves for an high arch
dam in which both the material and ground motion uncertainties were
considered. Limit states were selected based on the concrete damage
and amount of joint opening. Analyses were performed for 10 three-
component ground motions, each one scaled to 10 intensity levels.
Fragility curves were developed using both the log-normal CDF ans
spline fitting. Hariri-Ardebili et al. [119] [ID-29] proposed a response
surface meta-model for material uncertainty quantification of an high
arch dam subjected to a strong ground shaking. The model is based on
the Box-Behnken Experimental Design where the concrete elastic
modulus, mass density, tensile and compressive strength are assumed to
be random variables. Several surrogate models are developed for the
maximum crest displacement, concrete cracking and crushing. The
fragility curves are then presented based on both the empirical and log-
normal CDF functions.
4.5. Vulnerability
Vulnerability originates from the Latin word vulnerare, which il-
lustrates the potential to be harmed and the sensitivity to a perturbation
or stress [313]. Lei et al. [314] collected a comprehensive list of defi-
nitions related to vulnerability. Some of those are reported here:
822
International Journal of Disaster Risk Reduction 31 (2018) 806–831
• Vulnerability means an environmental sensitivity. It is related to
demographic, economic, social, and technical factors [313].
• When a social or ecological system loses resilience, it becomes
vulnerable to change that previously could be absorbed. Thus, it is
the flip side of resilience [315].
• Vulnerability is the degree to which a system is susceptible to, or
unable to cope with, adverse effects of climate change [316].
• Vulnerability is the degree to which a system, subsystem, or system
component is likely to experience harm due to hazardous exposures,
either perturbations or stressors [317].
• Social vulnerability is a measure of both the sensitivity of a popu-
lation to natural hazards and its ability to respond to and recover
from the impacts of hazards [318].
• Vulnerability presents the characteristics of a group or individual in
terms of its capacity to anticipate, cope with, resist, and recover
from the impact of a hazard [319].
• Vulnerability places stress on a system's response to (potential) ha-
zards, which determines the likelihood of loss. Exposure and sensi-
tivity are two aspects of vulnerability [320].
• Vulnerability is the result of interactive change of both exposure and
response, or the covariance between them [321].
In the field of earthquake engineering, vulnerability is different
from fragility. The former measures loss (in terms dollars, deaths, and
downtime), while the latter measures probability [322]. A vulnerability
curve expresses loss as a function of IM. Three major types of vulner-
ability curves are:
Measuring repair cost: In such a case, the repair cost is normalized
by the replacement cost of a new asset and is called a damage factor.
The expected value of a damage factor conditioned on an IM parameter
is called a mean damage factor, Fig. 18.
Measuring life safety: In such a case, the number of casualties is
normalized by the number of indoor occupants and expressed as a
function of IM parameter.
Measuring downtime: It is measured in terms of the fraction of a
year during which the structure cannot be used.
In general, there are three approaches to develop a vulnerability
function, i.e., empirical, analytical, and expert opinion [322]. An em-
pirical vulnerability function is derived by regression analysis of ob-
servations of pairs of excitation and loss of specimens of an asset class
[323]. This method is of interest to insurance companies, as the data
comes directly from real-world events. Analytical vulnerability func-
tions are useful in cases where there is little or no data from empirical
sources. Under these condition, with the time and knowledge to create
and analyze an analytical model, multiple numerical simulations are
performed. This method involves defining one or more specimens to
represent the class, creating and analyzing a structural model to esti-
mate demand parameters as a function of ground motion intensity
measure, estimating damage at component level, and finally estimating
loss given component damage. Lastly, in cases where empirical data
points are missing and performing analytical models are difficult (im-
possible or expensive), expert opinion can be used to develop
Fig. 18. Vulnerability function.
M.A. Hariri-Ardebili
vulnerability curves [286].
The following is a short review of the published papers in vulner-
ability analysis of dams. Baecher et al. [324] proposed a cost-benefit
analysis for dams that included the economic costs from dam failure
based on the net project benefits. Economic failure costs include
property damage, loss of future benefits, the lost economic activity, and
emergency costs. Failure costs are temporal in response to economic
development. Based on a reported example, they claimed that about
half of all dam failures occur 0–5 years after construction, while the
other half is uniformly distributed over the age of the dam. Paté-Cornell
and Tagaras [325] discuss cost-benefit analyses for a case in which a
new dam is going to be built in sequence with an existing dam. In such a
case, the marginal risk cost associated with the new dam should be
considered, including its interaction with other structures. They show
that risk costs might have an important role in the project's cost-benefit
ratio when the dam has a large floodplain, and the project costs and
benefits are relatively small.
Lave et al. [326] proposed an alternative safety goal for dams by
separating property damage from possible loss of life. For a potential
dam failure with large loss of life or property damage, a careful eva-
luation should be done as to whether the dam should be built. For dams
that impose smaller hazards, property damage should be analyzed
based on expected values of annual benefits and costs. Ellingwood et al.
[327] developed a methodology for estimating the total costs of a dam
failure. It takes into consideration the fundamental issues of accounting
stance, damage to fixed assets, loss of income, intangible losses, and loss
of life and limb. Cost-related information was portrayed by type and
collection priority in a data-collection matrix to guide an on-site in-
vestigation.
Ansar et al. [328] provided a unique paper on the overall actual cost
of dams and hydropower systems. They used a large data set in con-
junction with the “outside view” or “reference class forecasting” tech-
nique for the decision-making under uncertainty. They claim that, in
most countries, large hydropower dams will be too costly in absolute
terms and take too long to build to deliver a positive risk-adjusted re-
turn unless suitable risk management measures are used. They advise
policymakers to prefer agile energy alternatives that can be built over
shorter time lines. Morales-Torres et al. [329] provided a review of the
existing indicators for dam safety, tracking how equity and efficiency
principles are captured and proposing additional indicators on how
tolerability guidelines and cost-benefit analyses can be used in decision-
making. Pisaniello and Tingey-Holyoak [330] studied the inter-related
policy, responsibility, cost-sharing, and engineering issues associated
with dam safety to mitigate failure threats for both existing and future
downstream developments. It was based on 1) a review of Australian
practices underpinned by relevant theoretical principles, and 2) the
development of a cost-effective flood safety review/design tool to help
policy makers address cost-sharing issues.
4.6. Resilience
The Presidential Policy Directive [331] defines the term resilience
as “a structure's ability to prepare for and adapt to changing conditions
while simultaneously being able to withstand and recover rapidly from
functionality disruptions.” There are several analytical studies that
quantify the resiliency in civil infrastructures subjected to an extreme
event discussed in [332–336]. Resilience metrics for multi-hazard en-
vironments were defined by Ayyub [337]. The Task Committee (TC) on
Infrastructure Impacts of Landscape-Driven Weather Change of the
ASCE provided the following definition [338]: “A weather-climate re-
silient water infrastructure is defined as an infrastructure that can to a
degree anticipate or adapt and recover from external disruptions due to
severe weather and climate, and carry on providing the essential ser-
vices the infrastructure is designed for with managed interruption to
823
International Journal of Disaster Risk Reduction 31 (2018) 806–831
non-essential services, while balancing trade-offs among social (e.g.,
security), environmental, and economic factors.”
Bruneau et al. [339] described one of the most comprehensive de-
finitions of resilience that accounts for eleven aspects categorized in
three groups [334]:
• Four dimensions:
Technical: includes all the technological aspects associated with the
construction.
Organizational: includes all the management activities and plans,
maintenance, and response to emergencies.
Social: includes societal effects and its mitigation.
Economic: includes all the indirect and direct costs due to the loss
of functionality and the subsequent rehabilitation.
• Four properties:
Robustness: the ability to withstand a given extreme event and still
deliver a service.
Rapidity: the speed with which a structure recovers from such an
event to reach a high functionality level.
Redundancy: the extent to which components of the system are
substitutable.
Resourcefulness: the capacity to make the appropriate budget
available, identify problems, establish priorities, and mobilize re-
sources after an extreme event.
• Three results:
More reliable: since it has a lower probability of reaching limit
states.
Fast recovery: the rapidity of functionality restoration during a
disaster is paramount to resilient systems.
Low socio-economic consequences: the reduced probability of
significant service reductions and fast recovery decreases the impact
of extreme events on a society.
Community resilience is the ability to prepare and plan for, absorb,
recover from, and more successfully adapt to actual or potential adverse
events [340]. Resilience is defined as a normalized function, indicating
capability to sustain a level of functionality or performance, Q (t ) , for a
given building, lifeline, or community over a period of time, tLC , (life
cycle time) [341]. tLC includes the structure recovery time, tRE , and the
business interruption time, tBI (usually negligible). tRE is the time ne-
cessary to restore the functionality of a critical infrastructure system
(and usually is a RV with high uncertainties). Resilience can be defined
as [342]:
Fig. 19. Resilience curve.
M.A. Hariri-Ardebili
tRE Q (t )
Resilience (%) = ∫t
0 tRE
dt
(36)
where t0 is the hazard (e.g., earthquake) occurrence time. Resilience
and loss of resilience, its complement, are usually shown through a
“recovery function”, as shown in Fig. 19.
Furthermore, resilience and risk are connected to each other.
Improving the resilience of a structural system constitutes an integral
part of its risk management [343]. More specifically, the acceptable
level of residual risk (e.g. recovery time and composite costs) in the
system can be quantified, subjected to a particular threat scenario, and
presented based on the associated PDFs. Manyena and Gordon [344]
compared and contrasted the concept fo resilience with fragility and
stabilization. In the bigger picture, the major focus of resilience is on
agencies, while it is the structures in fragility. The delivery of resilience
is bottom-up, while it is top-down in fragility and stabilization. Simi-
larly, Corotis [345] discusses the conceptual and analytical differences
between resiliency and reliability.
There are few studies on the resilience of dams/levees. Graham
[346] discusses the potential structural retrofitting and upgrading of
dams from the economic and loss of life perspectives. The proposed
method (which in fact was developed to avoid costly over-designing)
provided logic for addressing the concept of resilience. The method was
based on calculating the following steps for each modification design
[347]: 1) annual cost of the modification, 2) annual economic loss
caused by flooding, 3) annual loss of life due to flooding, 4) loss of life
from construction spending (0.14 lives per $100 million expended) and
convert to annual value, 5) economic benefits derived from modifica-
tion, 6) life benefits derived from modification, 7) acceptance or re-
jection of modifications should be based on Table-2 of Graham [346].
The report published by National Research Council [348] might be
the most comprehensive official source of dam/levee resilience. In the
context of this report, “community” includes all persons and organi-
zations exposed to direct and indirect consequences of dam or levee
Fig. 20. Framework for resilience-based collaboration in dam/levee safety; adopted from National Research Council [348].
824
International Journal of Disaster Risk Reduction 31 (2018) 806–831
failure. The availability of hazard- and risk-related information is im-
portant for informed decision-making on the dam professionals com-
munity. Resilience-focused collaborations are a means of increasing
understanding and communication of risks, shared needs, and oppor-
tunities. Social capital is necessary for community resilience. However,
in majority of the cases, the dam and levee safety programs operate
independently of other community functions. Risk-informed ap-
proaches allow dam and levee professionals to improve their under-
standings of infrastructure-system operation, performance, vulner-
ability, and the consequences of potential failures. Finally, the dam
safety program needs a cultural shift to be able to properly account for
the community resilience. Fig. 20 illustrates a conceptual framework for
resilience-based collaboration in dam/levee safety. Enhancing resi-
lience requires frequent and collaborative assessment of risk. The Na-
tional Research Council [348] proposed a maturity matrix for assessing
community engagement in dam safety programs. This is a general fra-
mework, which should be adjusted on a case-by-case basis. It uses a
matrix to assess how advanced a program is with respect to a specific
goal. The rows describe specific program or community processes,
while columns refer to tools that are in place at given increased levels of
community engagement related to that function. Table 3 shows a
sample maturity matrix.
Kim et al. [349] proposed a framework for evaluating dam safety
strengthening alternatives in terms of a comprehensive flood control by
applying the concept of resilience. Karamouz et al. [350] studied the
regional drought resiliency in a specific location, which included a
major dam. The multi-criteria, decision-making approach was used to
determine the relative importance of the four resiliency components.
They found that the downstream area of the case study dam is more
resilient compared to the upstream one. Combining the robustness and
the cost indices, Lee et al. [351] suggested a hybrid index as an in-
dicator of the robustness of the structural alternatives for flood control
purposes.
M.A. Hariri-Ardebili International Journal of Disaster Risk Reduction 31 (2018) 806–831

Table 3
Sample maturity matrix for assessing community engagement; adopted from National Research Council [348].
Elements Level (I) Level (II) Level (III) Level (IV) Level (V)

Dam safety reviews No activity Standards-based only Introduction of additional Application of quantitative risk Application of quantitative
review criteria (e.g., failure assessment by using criteria risk assessment by using
mode analysis) developed by owner or regulator with criteria that reflect the
input from community members and community's societal values
stakeholders
Emergency action No activity EAPs developed EAPs developed with input EAPs developed with input from Community collaboration
plans internally by owner from emergency management community members and with owners or operators to
agency stakeholders and emergency develop integrated EAPs that
management agency and shared with reflect community values
selected community representatives
Floodplain No floodplain Floodplain management Floodplain management Floodplain management plans Floodplain management
management management plans plans in place plans accommodate shadow integrated into community plans fully integrated into
floodplain associated with comprehensive or general plans dam owners’ planning
catastrophic dam failure processes

Bheemasetti et al. [352] presented a framework for sustainability
and resiliency and how they could be used to assess infrastructures. A
case study of aging levees was conducted to optimize the number of in-
situ seismic cone penetration tests to provide a balanced, en-
vironmentally-friendly, and cost-effective field exploration, while still
providing reliable soil layer configurations needed for resiliency ana-
lyses. Moreover, Kim et al. [353] applied the concept of resilience to
evaluate dam safety upgrade options from a flood damage mitigation
perspective. They defined resilience as the ability to provide flood-
control service after an extreme rainfall event and to recover a desired
level of functionality as soon as practically available. Their framework
includes four steps: 1) estimating the probable maximum flood, 2)
calculating the dam discharge for each of the upgrade options, 3)
quantifying the loss in system resilience, and 4) converting the losses
into monetary units. Finally, Matthews and McCartney [354] explored
the challenges facing decision makers with regard to building resilience
and navigating risks within the water-energy-food nexus and dams.
Two case studies in Africa and Asia were conducted, and the authors
underlined the need for extra efforts in this field.
4.7. Sustainability
Sustainability is another concept which is similar to resilience but
has its own definition. While resilience is presented as a metric to
measures the ability of a structural system to withstand and recover
from an unusual event, sustainability is a model that addresses si-
multaneously the current needs and the impacts on future generations
[355]. Sustainability combines three dimensions: economic, environ-
mental, and social [334]. Different aspects of sustainability can be re-
presented in four levels (associated with project management), as
shown in Fig. 21, where the lower levels influence some goals asso-
ciated with the higher ones. Usually sustainability is associated with the
words like “green design”, “life cycle”, and “social cost”, whereas re-
silience is associated with “recovery”, extreme events” and “ function-
ality”. In spite of the resilience (which is in the community and network
level), most of the current applications in sustainability are limited to
the buildings and bridges [356,357]. Furthermore, Marchese et al.
[358] compared and contrasted the resilience and sustainability of
environmental management applications. They considered three sce-
narios: 1) resilience as a component of sustainability, 2) sustainability
as a component of resilience, and 3) resilience and sustainability as
separate objectives.
Studies on the sustainability of the dams and levees are very rare.
Basu et al. [359] provided an overview of the sustainability within the
field of geotechnical engineering, including some insights on dams,
embankments, and levees. They emphasized the importance of the
sustainability and the fact that the current practices do not put much
value in the sustainability of geo-structures. Das et al. [360] compared
sustainability and resilience for a slope stability project. More specifi-
cally, they quantified these concepts by assigning appropriate weights,
ωi , showing the relative importance of discretized items. The sustain-
ability index was presented as:
Isus = ω1 Irec + ω2 Ienv + ω3 Iso / eco (37)

Fig. 21. Hierarchy levels in sustainability; adopted from Bocchini et al. [334].

825
M.A. Hariri-Ardebili
where Irec , Ienv , and Iso / eco are the resource, environmental, and socio/
economic impacts, respectively. They also combined Isus with the resi-
lience index, Ires , to create a cumulative sustainability and resilience
index, ICSR = ω4 Isus + ω5 Ires . Subsequently, treatment methods with a
minimum ICSR is recommended.
Sevionovic [361] quantified risk as one of the key sustainability
issues in water resources. They evaluated the risk based on the prob-
abilistic and fuzzy set theory and extended it to risk management.
Yüksel [362] discussed the policies needed to meet increasing energy
demands for sustainable hydro-power energy development in Turkey.
From a sustainable management point of view, Yang et al. [363] as-
sessed several alternatives for managing a small flood-control dam,
including improving dam safety, controlling floods, and restoring ri-
parian vegetation. Bird [364] studied the socio-economic impact of
hydro-electric dams on developing communities through a case study of
the Chalillo Dam on the Macal River in Belize. Quantitative data was
collected using a set of questionnaires with a likert response scale.
Tessler et al. [365] quantified variations in flood risks associated
with extreme events based on an integrated set of global environmental,
geophysical, and social indicators in which the risk is distributed across
all levels of economic development. Moreover, Wieland [366] ex-
plained the concept of a sustainable dam in terms of dam safety and
environmental-economic-social aspects. He explained that the non-
technical features of the dam (i.e. environmental, economic, social) are
important to ensure safety is guaranteed. For this reason, when de-
signing or constructing a sustainable (multi-purpose) dam, the project
manager must be a dam or civil engineer and not a person with in-
adequate technical background experience.
Acknowledgment
The author would like to express his sincere appreciation to his
former advisor (and the current PI), Prof. Victor E. Saouma at the
University of Colorado Boulder for his enthusiastic guidance and advice
throughout this research. Moreover, the author would like to thank
Prof. Upmanu Lall at the Columbia University and director of Columbia
Water Center, for his helpful and constructive comments, which have
greatly contributed to improving the final version of this paper. General
discussion by Dr. Jerzy W. Salamon at US Bureau of Reclamation (chair
of Concrete Committee at the US Society of Dams), and Mr. Larry K.
Nuss (ret. from USBR and Nuss Engineering LLC) is highly appreciated.
Last but not least, the author highly appreciate the positive and helpful
comments provided by two anonymous reviewers.
References
[1] L. Zhang, Social impacts of large dams: the China case, Citeseer (2000).
[2] SPANCOLD, Risk analysis applied to management of dam safety, Technical guide
on operation of dams and reservoirs, Volume 1, Techical Report. in: Proceedings of
the Spanish National Committee on Large Dams, Madrid, Spain, 2013.
[3] ASCE, American Society of Civil Engineers, 〈https://www.
infrastructurereportcard.org/cat-item/dams/〉, Note = Last viewed January 2018,
2017 Report Card for America’s Infrastructure; Dams.
[4] FEMA, Federal Emergency Management Agency, 〈https://www.fema.gov/media-
library-data/20130726-1737-25045-8253/1_2010esri_damsafety061711.pdf〉,
Note = Last viewed January 2018, Identifying High Hazard Dam Risk in the
United States, year = 2014.
[5] FERC-Arch, Engineering guidelines for the evaluation of hydropower projects:
Chapter 11: Arch dams, Technical Report. in: Proceedings of the Federal Energy
Regulatory Commission, Washington, DC, USA, 1999.
[6] USBR-manual, Dam safety risk analysis, Best practices training manual, Version 2.
2, Tech. Rep., U.S. Department of the Interior Bureau of Reclamation in co-
orporation with the U.S. Army Corps of Engineers, Denver, Colorado, 2011.
[7] ANCOLD, Guidelines on risk assessment, Technical Report. in: Proceedings of the
Australian National Committee on Large Dams, Sydney, New South Wales,
Australia, 2003.
[8] Canadian Dam Association (CDA), Dam safety guidelines, Technical Report.
Edmonton, Alberta, Canada, 2007.
[9] Y. Ghanaat, Failure modes approach to safety evaluation of dams, in: Proceedings of the
13th World Conference on Earthquake Engineering, Vancouver, BC, Canada, 2004.
[10] L. Zhang, M. Peng, D. Chang, Y. Xu, Dam failure mechanisms and risk assessment,
826
International Journal of Disaster Risk Reduction 31 (2018) 806–831
John Wiley & Sons, 2016.
[11] V. Singh, Dam Breach Model. Technol. 17 (1996).
[12] A.T.C. on Dam/Levee Breaching, Earthen embankment breaching, Journal of hy-
draulic engineering 137 (12) (2011) 1549–1564.
[13] L. Richardson, The approximate arithmetical solution by finite differences of
physical problems involving differential equations, with an application to the
stresses in a masonry dam, Philos. Trans. R Soc. Ser. A 210 (1911) 307–357.
[14] R. Clough, E. Wilson, Stress analysis of a gravity dam by the finite element
method, in: Proceedings of the Symposium on Use of Computers in Civil
Engineering, Lisbon, Portugal, 1962.
[15] E. Bretas, J. Lemos, P. Lourenço, A {DEM} based tool for the safety analysis of
masonry gravity dams, Eng. Struct. 59 (2014) 248–260.
[16] E. Oñate, M.A. Celigueta, S.R. Idelsohn, F. Salazar, B. Suárez, Possibilities of the
particle finite element method for fluid-soil-structure interaction problems,
Comput. Mech. 48 (3) (2011) 307.
[17] R. Das, P. Cleary, A mesh-free approach for fracture modelling of gravity dams
under earthquake, Int. J. Fract. 179 (2013) 9–33.
[18] S. Zhang, G. Wang, X. Yu, Seismic cracking analysis of concrete gravity dams with
initial cracks using the extended finite element method, Eng. Struct. 56 (2013)
528–543.
[19] M. Goldgruber, R. Malm, Nonlinear Seismic Simulation of an Arch Dam using
XFEM, in: Simulia Regional Users Meeting, Graz, Austria., 2014.
[20] M. Abdollahi, R. Attarnejad, Dynamic analysis of dam-reservoir-foundation in-
teraction using finite difference technique, J. Cent. South Univ. 19 (2012)
1399–1410.
[21] M.A. Hariri-Ardebili, M.R. Kianoush, Integrative seismic safety evaluation of a
high concrete arch dam, Soil Dyn. Earthq. Eng. 67 (2014) 85–101.
[22] ANSYS, ANSYS software reference manuals, release notes, mechanical APDL, elements
reference, commands reference and theory reference, version Release 11, 2007.
[23] ABAQUS, Abaqus Theory Manual, ABAQUS, Inc., Providence, RI, USA, version6.7
edn., 2007.
[24] V. Saouma, J. Červenka, R. Reich, Merlin Finite Element User’s Manual, 〈http://
civil.colorado.edu/saouma/pdf/Software/users.pdf〉, 2010.
[25] F. Salazar, J. San Mauro, D. Vicente, C. Baena, C. Granell, L. Gracia, I. de
Pouplana, E. Oñate, Computer-aided design and analysis of arch dams.
[26] H. Westergaard, Water pressures on dams during earthquakes, Trans. Am. Soc.
Civil. Eng. 98 (1933) 418–433.
[27] C. Zangar, Hydrodynamic pressures on dams due to horizontal earthquakes, Proc.
Soc. Exp. Stress Anal. 10 (1953) 93–102.
[28] A.K. Chopra, Reservoir-dam interaction during earthquakes, Bull. Seismol. Soc.
Am. 57 (4) (1967) 675–687.
[29] P. Chakrabarti, A. Chopra, Hydrodynamic effects in earthquake response of gravity
dams, J. Struct. Div. 100 (6) (1974) 1211–1224.
[30] A. Chopra, P. Chakrabarti, Earthquake analysis of concrete gravity dams including
dam-water-foundation rock interaction, Earthq. Eng. Struct. Dyn. 9 (1981)
363–383.
[31] G. Fenves, A. Chopra, Earthquake analysis of concrete gravity dams including
reservoir bottom absorption and dam-water-foundation rock interaction, Earthq.
Eng. Struct. Dyn. 12 (1984) 663–680.
[32] G. Fenves, A.K. Chopra, Simplified earthquake analysis of concrete gravity dams,
J. Struct. Eng. 113 (8) (1987) 1688–1708.
[33] M. Yazdchi, N. Khalili, S. Valliappan, Dynamic soil-structure interaction analysis
via coupled finite-element-boundary-element method, Soil Dyn. Earthq. Eng. 18
(7) (1999) 499–517.
[34] N. Bouaanani, F. Lu, Assessment of potential-based fluid finite elements for seismic
analysis of dam-reservoir systems, Comput. Struct. 87 (2009) 206–224.
[35] G. Hou, J. Wang, A. Layton, Numerical methods for fluid-structure interaction - a
review, Commun. Comput. Phys. 12 (2) (2012) 337–377.
[36] B. Miquel, N. Bouaanani, Accounting for earthquake-induced dam-reservoir in-
teraction using modified accelerograms, J. Struct. Eng. 139 (9) (2013) 1608–1617.
[37] N. Bouaanani, S. Renaud, Effects of fluid-structure interaction modeling assump-
tions on seismic floor acceleration demands within gravity dams, Eng. Struct. 67
(2014) 1–18.
[38] U. Basu, A. Chopra, Perfectly matched layers for transient elastodynamics of un-
bounded domains, Int. J. Numer. Methods Eng. 59 (2004) 1039–1074.
[39] V. Saouma, F. Miura, G. Lebon, Y. Yagome, A simplified 3D model for soil-struc-
ture interaction with radiation damping and free field input, Bull. Earthq. Eng. 9
(2011) 1387–1402.
[40] A. Løkke, A.K. Chopra, Direct finite element method for nonlinear analysis of semi-
unbounded dam-water-foundation rock systems, Earthq. Eng. Struct. Dyn. 46 (8)
(2017) 1267–1285.
[41] M.K. Poul, A. Zerva, Nonlinear dynamic response of concrete gravity dams con-
sidering the deconvolution process, Soil Dyn. Earthq. Eng. 109 (2018) 324–338.
[42] K. Willam, E. Pramono, S. Stur, Fundamental issues of smeared crack models, in:
Shah, Swartz, , (Eds.), in: Proceedings of the SEM-RILEM International Conference
on Fracture of Concrete and Rock. 1987 pp.192–207.
[43] L. Malvar, M. Fourney, A three dimensional application of the smeared crack
approach, Eng. Fract. Mech. 35 (1990) 251–260.
[44] S. Weihe, B. Kroplin, R. DeBorst, Classification of smeared crack models based on
material and structural properties, Int. J. Solids Struct. 35 (1998) 1289–1308.
[45] J. Lee, G.L. Fenves, A plastic-damage concrete model for earthquake analysis of
dams, Earthq. Eng. Struct. Dyn. 27 (9) (1998) 937–956.
[46] J. Moslera, G. Meschke, Computational failure mechanics embedded crack vs.
smeared crack models: a comparison of element wise discontinuous crack path
approaches with emphasis on mesh bias, Comput. Methods Appl. Mech. Eng. 193
(2004) 3351–3375.
M.A. Hariri-Ardebili
[47] M. Cervera, M. Chiumenti, Smeared crack approach: back to the original track, Int.
J. Numer. Anal. Methods Geomech. 30 (2006) 1173–1199.
[48] G. Fenves, S. Mojtahedi, R. Reimer, Parameter study of joint opening effects on
earthquake response of arch dams, Technical Report No. 92/05, Tech. Rep.,
Earthquake Engineering Research Center, University of California at Berkeley,
Berkeley, California, US, 1992.
[49] J. Hall, Efficient non-linear seismic analysis of arch dams, Earthq. Eng. Struct.
Dyn. 27 (1998) 1425–1444.
[50] M. Ahmadi, M. Izadinia, H. Bachmann, A discrete crack joint model for nonlinear
dynamic analysis of concrete arch dam, Comput. Struct. 79 (2001) 403–420.
[51] M. Azmi, P. Paultre, Three-dimensional analysis of concrete dams including con-
traction joint non-linearity, Eng. Struct. 24 (6) (2002) 757–771.
[52] E. Puntel, G. Bolzon, V. Saouma, Fracture mechanics based model for joints under
cyclic loading, ASCE J. Eng. Mech. 132 (2006) 1151–1159.
[53] C. Zhang, Y. Xu, G. Wang, F. Jin, Non-linear seismic response of arch dams with
contraction joint opening and joint reinforcements, Earthq. Eng. Struct. Dyn. 29
(10) (2000) 1547–1566.
[54] H. Zhong, H. Li, Y. Bao, Seismic Risk Analysis of an Arch Dam, in: Applied
Mechanics and Materials, vol. 353-356, Trans Tech Publ, 2020–2023, 2013.
[55] Y. Long, C. Zhang, Y. Xu, Nonlinear seismic analyses of a high gravity dam with
and without the presence of reinforcement, in: Seismic Safety Evaluation of
Concrete Dams, Elsevier, 369-386, 2014.
[56] M. Wang, J. Chen, S. Fan, S. Lv, Experimental study on high gravity dam
strengthened with reinforcement for seismic resistance on shaking table, Struct.
Eng. Mech. 51 (2014) 663–683.
[57] R. Malm, Guideline for FE analyses of concrete dams, Tech. Rep., Energiforsk
Report, Sweden, 2016.
[58] FEMA-PFM, Selecting analytic tools for concrete dams to address key events along
potential failure mode paths, Technical Report. Federal Emergency Management
Agency, Denver, Colorado, 2011.
[59] M.T. Schultz, B.P. Gouldby, J.D. Simm, J.L. Wibowo, Beyond the factor of safety:
Developing fragility curves to characterize system reliability, Techical Report.
DTIC Document, 2010.
[60] G. Ruggeri, Sliding Safety of Existing Gravity Dams, Final Report, ICOLD European
Club.
[61] J. Spross, A Critical Review of the Observational Method (Ph.D. thesis), KTH Royal
Institute of Technology, 2014.
[62] L. Altarejos-García, I. Escuder-Bueno, A. Serrano-Lombillo, A. Morales-Torres,
Factor of safety and probability of failure in concrete dams, Risk Anal. Dam Saf.
Dam Secur. Crit. Infrastruct. Manag. (2012) 251–256.
[63] A. Krounis, Uncertainty in Sliding Stability Analyses of Existing Concrete Gravity
Dams with Bonded Concrete-Rock Interfaces (Ph.D. Thesis), KTH Royal Institute of
Technology, 2013.
[64] M.A. Hariri-Ardebili, F. Pourkamali-Anaraki, Support vector machine based re-
liability analysis of concrete dams, Soil Dyn. Earthq. Eng. 104 (2018) 276–295.
[65] F. Silva, T.W. Lambe, W.A. Marr, Probability and risk of slope failure, J. Geotech.
Geoenviron. Eng. 134 (12) (2008) 1691–1699.
[66] A. Yudelevich, Uncertainty of initial information in reliability assessment of
concrete dams (within the framework of structural theory), Power Technol. Eng.
51 (5) (2018) 545–551.
[67] A. Der-Kiureghian, O. Ditlevsen, Aleatory or epistemic? Does it matter? Struct. Saf.
31 (2009) 105–112.
[68] B. Möller, W. Graf, M. Beer, Safety assessment of structures in view of fuzzy
randomness, Comput. Struct. 81 (15) (2003) 1567–1582.
[69] G.B. Baecher, Uncertainty in dam safety risk analysis, Georisk: Assess. Manag. Risk
Eng. Syst. Geohazards 10 (2) (2016) 92–108.
[70] Army Corps of Engineers, Reliability Assessment of Navigation Structures, ETL 1110-2-
532, Department of the Army, US Army Corps of Engineers, Washington, D.C., 1992.
[71] A. Yücel, Seismic Analysis of Concrete Gravity Dams Including Dam-Foundation-
Reservoir Interaction, Master’s thesis, Middel East technical University, Turkey, 2013.
[72] M.A. Hariri-Ardebili, V.E. Saouma, Sensitivity and uncertainty quantification of
the cohesive crack model, Eng. Fract. Mech. 155 (2016) 18–35.
[73] R. Iman, W. Conover, A distribution-free approach to inducing rank correlation
among input variables, Commun. Stat.-Simul. Comput. 11 (3) (1982) 311–334.
[74] J.H. Halton, Algorithm 247: radical-inverse quasi-random point sequence,
Commun. ACM 7 (12) (1964) 701–702.
[75] I.M. Sobol', On the distribution of points in a cube and the approximate evaluation
of integrals, Zh. Vychislitel'noi Mat. i Mat. Fiz. 7 (4) (1967) 784–802.
[76] M. Cavazzuti, Optimization Methods: From Theory to Design Scientific and
Technological Aspects in Mechanics, Springer Science & Business Media, 2012.
[77] L. Faravelli, Response-surface approach for reliability analysis, J. Eng. Mech. 115
(12) (1989) 2763–2781.
[78] M.A. Hariri-Ardebili, MCS-based response surface metamodels and optimal design
of experiments for gravity dams, Struct. Infrastruct. Eng. (2018) 1–23.
[79] M. Hariri-Ardebili, P. Boodagh, Taguchi design-based seismic reliability analysis of
geostructures, Georisk: Assess. Manag. Risk Eng. Syst. Geohazards (2018) 1–19.
[80] M.A. Hariri-Ardebili, S.M. Seyed-Kolbadi, M. Noori, Response surface method for
material uncertainty quantification of infrastructures, Shock Vib. (2018).
[81] J. Brownlee, A Tour of Machine Learning Algorithms, URL 〈https://
machinelearningmastery.com/a-tour-of-machine-learning-algorithms/〉, 2013.
[82] F. Salazar, M. Toledo, E. Oñate, R. Morán, An empirical comparison of machine
learning techniques for dam behaviour modelling, Struct. Saf. 56 (2015) 9–17.
[83] F. Salazar, R. Morán, M.Á. Toledo, E. Oñate, Data-based models for the prediction
of dam behaviour: a review and some methodological considerations, Arch.
Comput. Methods Eng. (2015) 1–21.
[84] V. Saouma, E. Hansen, B. Rajagopalan, Statistical and 3d nonlinear finite element
827
International Journal of Disaster Risk Reduction 31 (2018) 806–831
analysis of Schlegeis dam, in: Proceedings of the Sixth ICOLD Benchmark
Workshop on Numerical Analysis of Dams. 2001. pp.17–19.
[85] A. Gaspar, F. Lopez-Caballero, A. Modaressi-Farahmand-Razavi, A. Gomes-
Correia, Methodology for a probabilistic analysis of an RCC gravity dam con-
struction. Modelling of temperature, hydration degree and ageing degree fields,
Eng. Struct. 65 (2014) 99–110.
[86] J. Mata, N.S. Leitão, A.T. de Castro, J.S. da Costa, Construction of decision rules
for early detection of a developing concrete arch dam failure scenario. A dis-
criminant approach, Comput. Struct. 142 (2014) 45–53.
[87] H.-z. Su, Z.-p. Wen, C.-s. Gu, An early-warning model of dam safety based on
rough set theory and support vector machine, in: Proceedings of the 5th
International Conference on Machine Learning and Cybernetics, IEEE, Dalian,
China, 2006. pp. 3455–3460.
[88] O. Dancea, O. Tsatos, M. Gordan, A. Vlaicu, Adaptive fuzzy c-means through
support vector regression for segmentation of calcite deposits on concrete dam
walls, in: Proceedings of the International Conference on Automation Quality and
Testing Robotics (AQTR). vol. 3, IEEE, 2010. pp. 1–6.
[89] C. Gu, B. Li, G. Xu, H. Yu, Back analysis of mechanical parameters of roller
compacted concrete dam, Sci. China Technol. Sci. 53 (3) (2010) 848–853.
[90] H. Su, Z. Wen, S. Zhang, S. Tian, Method for choosing the optimal resource in
back-analysis for multiple material parameters of a dam and its foundation, J.
Comput. Civ. Eng. 30 (4) (2016).
[91] L. Cheng, D. Zheng, Two online dam safety monitoring models based on the
process of extracting environmental effect, Adv. Eng. Softw. 57 (2013) 48–56.
[92] V. Ranković, N. Grujović, D. Divac, N. Milivojević, Development of support vector
regression identification model for prediction of dam structural behaviour, Struct.
Saf. 48 (2014) 33–39.
[93] H. Su, Z. Chen, Z. Wen, Performance improvement method of support vector
machine-based model monitoring dam safety, Struct. Control Health Monit. 23 (2)
(2016) 252–266.
[94] Y. Yu, X. Liu, E. Wang, K. Fang, L. Huang, Dam safety evaluation based on multiple
linear regression and numerical simulation, Rock Mech. Rock. Eng. (2018) 1–17.
[95] L. Cheng, F. Tong, Y. Li, J. Yang, D. Zheng, Comparative study of the dynamic
back-analysis methods of concrete gravity dams based on multivariate machine
learning models, J. Earthq. Eng. (2018) 1–22.
[96] M. Alembagheri, M. Ghaemian, Seismic assessment of concrete gravity dams using
capacity estimation and damage indexes, Earthq. Eng. Struct. Dyn. 42 (2013)
123–144.
[97] M. Alembagheri, M. Ghaemian, Incremental dynamic analysis of concrete gravity
dams including base and lift joints, Earthq. Eng. Eng. Vib. 12 (1) (2013) 119–134.
[98] M. Alembagheri, M. Ghaemian, Damage assessment of a concrete arch dam
through nonlinear incremental dynamic analysis, Soil Dyn. Earthq. Eng. 44 (2013)
127–137.
[99] A. Amirpour, H. Mirzabozorg, Quantifying the qualitative limit-states using IDA
approach in concrete arch dams, Arab. J. Sci. Eng. 39 (11) (2014) 7729–7740.
[100] J. Pan, Y. Xu, F. Jin, Seismic performance assessment of arch dams using incre-
mental nonlinear dynamic analysis, Eur. J. Environ. Civ. Eng. 19 (3) (2015)
305–326.
[101] B. Soysal, B. Binici, Y. Arici, Investigation of the relationship of seismic intensity
measures and the accumulation of damage on concrete gravity dams using in-
cremental dynamic analysis, Earthq. Eng. Struct. Dyn. 45 (2016) 719–737.
[102] G. Wang, Y. Wang, W. Lu, C. Zhou, M. Chen, P. Yan, XFEM based seismic potential
failure mode analysis of concrete gravity dam-water-foundation systems through
incremental dynamic analysis, Eng. Struct. 98 (2015) 81–94.
[103] M. Alembagheri, M. Ghaemian, Seismic performance evaluation of a jointed arch
dam, Struct. Infrastruct. Eng. 12 (2) (2016) 256–274.
[104] F.B. Soysal, B.Ö. Ay, Y. Arici, An investigation of the ground motion scaling
procedures for the nonlinear seismic analyses of concrete gravity dams, J. Earthq.
Eng. (2017) 1–24.
[105] M. Hariri-Ardebili, V. Saouma, Single and multi-hazard capacity functions for
concrete dams, Soil Dyn. Earthq. Eng. 101 (2017) 234–249.
[106] M. Leclerc, P. Léger, R. Tinawi, Computer aided stability analysis of gravity dams
CADAM, Adv. Eng. Softw. 34 (7) (2003) 403–420.
[107] J.-T. Wang, A.-Y. Jin, X.-L. Du, M.-X. Wu, Scatter of dynamic response and damage
of an arch dam subjected to artificial earthquake accelerograms, Soil Dyn. Earthq.
Eng. 87 (2016) 93–100.
[108] I.D. Gupta, S.G. Joshi, Response spectrum-based stochastic method for earthquake
analysis of gravity dams, J. Eng. Mech. 143 (5) (2017) 04017007.
[109] Z. Liu, Z. Liu, D. Chen, Probability density evolution of a nonlinear concrete
gravity dam subjected to nonstationary seismic ground motion, J. Eng. Mech. 144
(1) (2017) 04017157.
[110] M.A. Hariri-Ardebili, V.E. Saouma, Quantitative failure metric for gravity dams,
Earthq. Eng. Struct. Dyn. 44 (2015) 461–480.
[111] M.A. Hariri-Ardebili, L. Furgani, M. Meghella, V.E. Saouma, A new class of seismic
damage and performance indices for arch dams via ETA method, Eng. Struct. 110
(2016) 145–160.
[112] M.I. Ansari, P. Agarwal, Damage index evaluation of concrete gravity dam based
on hysteresis behavior and stiffness degradation under cyclic loading, Int. J.
Struct. Stab. Dyn. 17 (01) (2017) 1750009.
[113] A. Dell'Agnese, B. Mazzorana, F. Comiti, P. Von Maravic, V. D'agostino, Assessing
the physical vulnerability of check dams through an empirical damage index, J.
Agric. Eng. 44 (1) (2013) 2.
[114] M. Bybordiani, Y. Arıcı, The use of 3D modeling for the prediction of the seismic
demands on the gravity dams, Earthq. Eng. Struct. Dyn. 46 (11) (2017)
1769–1789.
[115] M. Alembagheri, M. Seyedkazemi, Seismic performance sensitivity and
M.A. Hariri-Ardebili
uncertainty analysis of gravity dams, Earthq. Eng. Struct. Dyn. 44 (1) (2015)
41–58.
[116] M. Andreini, A. De Falco, G. Marmo, M. Mori, G. Sevieri, Modelling issues in the
structural analysis of existing concrete gravity dams, in: Proceedings of the 85th
ICOLD Annual Meeting. 2017. pp. 363–383.
[117] M.P. Khiavi, Investigation of seismic performance of concrete gravity dams using
probabilistic analysis, Gradevinar 69 (1) (2017) 21–29.
[118] M.P. Khiavi, Investigation of the effect of reservoir bottom absorption on seismic
performance of concrete gravity dams using sensitivity analysis, KSCE J. Civil.
Eng. 20 (5) (2016) 1977–1986.
[119] M. Hariri-Ardebili, S. Seyed-Kolbadi, M. Noori, Response surface method for ma-
terial uncertainty quantification of infra-structures, Shock Vib. (2018).
[120] X. Tang, Y. Zhou, C. Zhang, J. Shi, Study on the heterogeneity of concrete and its
failure behavior using the equivalent probabilistic model, J. Mater. Civ. Eng. 23
(4) (2010) 402–413.
[121] H. Zhong, G. Lin, X. Li, J. Li, Seismic failure modeling of concrete dams con-
sidering heterogeneity of concrete, Soil Dyn. Earthq. Eng. 31 (12) (2011)
1678–1689.
[122] X. Yin, J. Li, G. Lin, H. Zhong, Damage simulation of large-scale concrete structures
subjected to earthquake excitation based on an equivalent damage element method,
in: Proceedings of the 15th World Conference on Earthquake Engineering, 2012.
[123] M. Hariri-Ardebili, S. Seyed-Kolbadi, V. Saouma, J. Salamon, B. Rajagopalan,
Random finite element method for the seismic analysis of gravity dams, Eng.
Struct. 171 (2018) 405–420.
[124] R. Malm, T. Gasch, D. Eriksson, M. Hassanzadeh, Probabilistic analyses of crack
propagation in concrete dams: Part 1, Elforsk report, Elforsk, 2013, p. 78 13:65.
[125] T. Gasch, R. Malm, D. Eriksson, M. Hassanzadeh, Probabilistic analyses of crack
propagation in concrete dams: Part II, Elforsk report, Elforsk, 2013, p. 53 13:66.
[126] R. Pereira, A.L. Batista, L.C. Neves, Probabilistic model for the representation of
the reservoir water level of concrete dams during normal operation periods, Water
Resour. Manag. 32 (9) (2018) 3041–3052.
[127] E.J. Anthony, Risk, vulnerability, and resilience: an overview, invulnerable Child
(1987) 3–48.
[128] T. Aven, On some recent definitions and analysis frameworks for risk, vulner-
ability, and resilience, Risk Anal. 31 (4) (2011) 515–522.
[129] T. Aven, A unified framework for risk and vulnerability analysis covering both
safety and security, Reliab. Eng. Syst. Saf. 92 (6) (2007) 745–754.
[130] A.T. Murray, T.H. Grubesic, Overview of reliability and vulnerability in critical
infrastructure, Crit. Infrastruct. (2007) 1–8.
[131] T. Hashimoto, J.R. Stedinger, D.P. Loucks, Reliability, resiliency, and vulnerability
criteria for water resource system performance evaluation, Water Resour. Res. 18
(1) (1982) 14–20.
[132] I. Escuder-Bueno, G. Mazzà, A. Morales-Torres, J.T. Castillo-Rodríguez,
Computational Aspects of Dam Risk Analysis: findings and Challenges,
Engineering 2 (3) (2016) 319–324.
[133] S. Marelli, R. Schobi, B. Sudret, UQLab user manual - Structural Reliability, Tech.
Rep., Chair of Risk, Safety and Uncertainty Quantification, ETH Zurich, report
UQLab-V0.92-107, 2016.
[134] M.A. Hariri-Ardebili, F. Pourkamali-Anaraki, Simplified reliability analysis of
multi hazard risk in gravity dams via machine learning techniques, Arch. Civil.
Mech. Eng. 18 (2) (2018) 592–610.
[135] O. Ditlevsen, H.O. Madsen, Structural Reliability Methods, 178 Wiley, New York,
1996.
[136] B. Dewals, S. Erpicum, S. Detrembleur, P. Archambeau, M. Pirotton, Failure of
dams arranged in series or in complex, Nat. Hazards 56 (2011) 917–939.
[137] S. Huaizhi, H. Jiang, Z. Wen, Service life predicting of dam systems with correlated
failure modes, ASCE J. Perform. Constr. Facil. 27 (2013) 252–269.
[138] A.M. Hasofer, N.C. Lind, Exact and invariant second-moment code format, Journal
of the Engineering Mechanics division 100 (1) (1974) 111–121. ASCE.
[139] K. Breitung, Asymptotic approximations for multinormal integrals, J. Eng. Mech.
110 (3) (1984) 357–366.
[140] M. McKay, R. Beckman, W. Conover, A comparison of three methods for selecting
values of input variables in the analysis of output from a computer code,
Technometrics 21 (2) (1979) 239–245.
[141] A. Olsson, G. Sandberg, Latin hypercube sampling for stochastic finite element
analysis, J. Eng. Mech. 128 (1) (2002) 121–125.
[142] A.B. Owen, Controlling correlations in Latin hypercube samples, J. Am. Stat.
Assoc. 89 (428) (1994) 1517–1522.
[143] Ayyub, Bilal M., Lai, Kwan-Ling, Structural reliability assessment using latin hy-
percube sampling, in: Structural Safety and Reliability, ASCE, 1177–1184, 1989.
[144] A. Harbitz, An efficient sampling method for probability of failure calculation,
Struct. Saf. 3 (2) (1986) 109–115.
[145] C.G. Bucher, Adaptive sampling - an iterative fast Monte Carlo procedure, Struct.
Saf. 5 (2) (1988) 119–126.
[146] R. Melchers, Search-based importance sampling, Struct. Saf. 9 (2) (1990) 117–128.
[147] R. Melchers, Radial importance sampling for structural reliability, J. Eng. Mech.
116 (1) (1990) 189–203.
[148] Y. Ibrahim, Observations on applications of importance sampling in structural
reliability analysis, Struct. Saf. 9 (4) (1991) 269–281.
[149] S. Au, J.L. Beck, A new adaptive importance sampling scheme for reliability cal-
culations, Struct. Saf. 21 (2) (1999) 135–158.
[150] R.E. Melchers, Structural Reliability Analysis and Prediction, John Wiley & Son
Ltd, 1999.
[151] J.D. Prendergast, Probabilistic Concept for Gravity Dam Analysis., Technical
Report. Construction Engineering Research Lab (ARMY) Champaign IL, 1979.
[152] H. Kreuzer, K. Bury, A probability based evaluation of the safety and risk of
828
International Journal of Disaster Risk Reduction 31 (2018) 806–831
existing dams, in: Proceedings of the International Conference on Safety of Dams.
1984. pp. 23–28.
[153] K. Bury, H. Kreuzer, Assessing the failure probability of gravity dams, Int. Water
Power dam Constr. 37 (11) (1985) 46–50.
[154] R. Baylosis, R. Bennett, Safety Assessment of an Existing Concrete Gravity Dam, in:
Proceedings of the Fifth International Conference on Structural Safety and
Reliability (ICOSSAR '89), San Francisco, California, US, 1989. pp. 279–286.
[155] T. Horyna, Reliability Analysis of Base Sliding of Concrete Gravity Dams Subjected
to Earthquakes (Ph.D. thesis), University of British Columbia, 1999.
[156] J. Jeppsson, Reliability-Based Assessment Procedures For Existing Concrete
Structures (Ph.D. thesis), Lund University, 2003.
[157] M.R. Kazemi, Reliability based analysis and design of anchor retrofitted concrete
gravity dams, in: Proceedings of the 13th World Conference on Earthquake
Engineering, 2004.
[158] V. Saouma, Reliability based nonlinear fracture mechanics analysis of a concrete
dam; a simplified approach, Dam Eng. 16 (3) (2006) 219–241.
[159] C. Carvajal, L. Peyras, J.-P. Bécue, C. Varon, C. Bacconnet, D. Clergue, D. Boissier,
Towards a probabilistic assessment of structural safety of gravity dams, in:
Proceedings of the 14th German Dam Symposium, 7th European Club of
International Commission on Large Dams (ICOLD), Munich, Germany, 2007.
[160] C. Carvajal, L. Peyras, C. Bacconnet, J. Bécue, Probability modelling of shear
strength parameters of RCC gravity dams for reliability analysis of structural
safety, Eur. J. Environ. Civil. Eng. 13 (2009) 91–119.
[161] C. Carvajal, L. Peyras, C. Bacconnet, On the loading/shear strength coupling in the
probabilistic formulation of the limit-state in shear for gravity dams, Eur. J.
Environ. Civil. Eng. 14 (3) (2010) 283–301.
[162] L. Peyras, C. Carvajal, H. Felix, C. Bacconnet, P. Royet, J. Becue, D. Boissier,
Probability-based assessment of dam safety using combined risk analysis and re-
liability methods - application to hazards studies, Eur. J. Environ. Civ. Eng. 16
(2012) 795–817.
[163] M. Papadrakakis, V. Papadopoulos, N.D. Lagaros, J. Oliver, A.E. Huespe,
P. Sánchez, Vulnerability analysis of large concrete dams using the continuum
strong discontinuity approach and neural networks, Struct. Saf. 30 (3) (2008)
217–235.
[164] G. Piliounis, N.D. Lagaros, Reliability analysis of geostructures based on meta-
heuristic optimization, Appl. Soft Comput. 22 (2014) 544–565.
[165] M. Westberg, Reliability-Based Assessment of Concrete Dam Stability (Ph.D.
thesis), Lund University, 2010.
[166] M. Westberg Wilde, F. Johansson, System reliability of concrete dams with respect
to foundation stability: application to a spillway, J. Geotech. Geoenviron. Eng. 139
(2) (2013) 308–319.
[167] Q. Xu, J. Li, J. Chen, Probability analysis for the damage of gravity dam,
Engineering 3 (4) (2011) 312–321.
[168] Q. Xu, J. Chen, J. Li, A study on the functional reliability of gravity dam, Energy
Power Eng. 4 (2) (2012) 59–66.
[169] Q. Xu, J.-Y. Chen, C. Zhao, J. Li, H. Yue, Failure probability analysis of interlayer
sliding belts of gravity dams under seismic load, Open Mech. Eng. J. 8 (2014)
85–91.
[170] M.E. Kartal, A. Bayraktar, H.B. Başağa, Seismic failure probability of concrete slab
on CFR dams with welded and friction contacts by response surface method, Soil
Dyn. Earthq. Eng. 30 (11) (2010) 1383–1399.
[171] M.E. Kartal, H.B. Başağa, A. Bayraktar, Probabilistic nonlinear analysis of CFR
dams by MCS using response surface method, Appl. Math. Model. 35 (6) (2011)
2752–2770.
[172] M.E. Kartal, A. Bayraktar, H.B. Başağa, Nonlinear finite element reliability analysis
of Concrete-Faced Rockfill (CFR) dams under static effects, Appl. Math. Model. 36
(11) (2012) 5229–5248.
[173] L. Altarejos-Garcia, I. Escuder-Bueno, A. Serrano-Lombillo, M. de Membrillera-
Ortuno, Methodology for estimating the probability of failure by sliding in con-
crete gravity dams in the context of risk analysis, Struct. Saf. 36–37 (2012) 1–13.
[174] A. Krounis, F. Johan, The Influence Of Correlation Between Cohesion and Friction
Angle on the Probability of Failure for Sliding of Concrete Dams, Risk Analysis,
Dam Safety, Dam Security and Critical Infrastructure Management, CRC Press,
2012, pp. 75–80.
[175] A. Krounis, F. Johansson, et al., A comparison between two techniques for in-
cluding the influence of progressive sliding failure in structural reliability analyses
of concrete dams, in: ISRM International Symposium-Proceedings of the 8th Asian
Rock Mechanics Symposium, International Society for Rock Mechanics, 2014.
[176] A. Krounis, F. Johansson, S. Larsson, Effects of spatial variation in cohesion over
the concrete-rock interface on dam sliding stability, J. Rock. Mech. Geotech. Eng.
7 (6) (2015) 659–667.
[177] A. Krounis, Sliding Stability Re-assessment of Concrete Dams with Bonded
Concrete-rock Interfaces, (Ph.D. thesis), KTH Royal Institute of Technology, 2016.
[178] H. Chen, W. Xu, Q. Wu, Z. Liu, S. Wang, Reliability analysis of arch dam subjected
to seismic loads, Arab. J. Sci. Eng. 39 (11) (2014) 7609–7619.
[179] C. Wang, S.-r. Zhang, B. Sun, G.-h. Wang, Methodology for estimating probability
of dynamical system's failure for concrete gravity dam, J. Cent. South Univ. 21
(2014) 775–789.
[180] G. Fiorentino, L. Furgani, S. Magliano, C. Nuti, Probabilistic evaluation of dam base
sliding, in: Proceedings of the 5th ECCOMAS Thematic Conference on Computational
Methods in Structural Dynamics and Earthquake Engineering, 2015.
[181] L. Altarejos-Garcia, I. Escuder-Bueno, A. Morales-Torres, Advances on the failure
analysis of the dam–foundation interface of concrete dams, Materials 8 (12)
(2015) 8255–8278.
[182] F. Liu, H. Li, Analysis of Time-dependent reliability of concrete gravity dam da-
mage based on gray theory, in: Proceedings of the 3rd International Conference on
M.A. Hariri-Ardebili
Management, Education, Information and Control (MEICI 2015), Shenyang,
China, 2015.
[183] Z. Shi, C. Gu, X. Zheng, D. Qin, Multiple failure modes analysis of the dam system
by means of line sampling simulation, Opt.-Int. J. Light Electron Opt. 127 (11)
(2016) 4710–4715.
[184] H. Su, J. Li, Z. Wen, Z. Fu, Dynamic non-probabilistic reliability evaluation and
service life prediction for arch dams considering time-varying effects, Appl. Math.
Model. 40 (15) (2016) 6908–6923.
[185] H. Chiti, M. Khatibinia, A. Akbarpour, H. Naseri, Reliability-based design opti-
mization of concrete gravity dams using subset simulation, Int. J. Optim. Civ. Eng.
6 (3) (2016) 329–348.
[186] G. Wang, Z. Ma, Application of probabilistic method to stability analysis of gravity
dam foundation over multiple sliding planes, Math. Probl. Eng. (2016).
[187] X. Zhu, X. Wang, X. Li, M. Liu, Z. Cheng, A new dam reliability analysis con-
sidering fluid structure interaction, Rock Mech. Rock Eng. (2017) 1–12.
[188] M. Cordier, P. Léger, Structural stability of gravity dams: a progressive assessment
considering uncertainties in shear strength parameters, Georisk Assess. Manag.
Risk Eng. Syst. Geohazards (2017) 1–14.
[189] M.A. Hariri-Ardebili, Analytical failure probability model for generic gravity dam
classes, Proc. Inst. Mech. Eng. Part O: J. Risk Reliab. 231 (5) (2017) 546–557.
[190] Y. Li, Y. Sun, B. Li, Z. Xu, Penalty function-based method for obtaining a reliability
indicator of gravity dam stability, Comput. Geotech. 81 (2017) 19–25.
[191] J. Jia, S. Wang, C. Zheng, Z. Chen, Y. Wang, FOSM-based shear reliability analysis
of CSGR dams using strength theory, Comput. Geotech. 97 (2018) 52–61.
[192] T.J. Ross, Fuzzy Logic with Engineering Applications, John Wiley & Sons, 2009.
[193] Q. Xu, J. Li, J.-y. Chen, Fuzzy reliability analysis of deep sliding plane in rock foun-
dation under dam, in: Proceedings of the Sixth International Conference on Fuzzy
Systems and Knowledge Discovery, 2009. FSKD'09. vol. 6, IEEE, 2009. pp. 525–529.
[194] C. Xin, G. Chongshi, Risk analysis of gravity dam instability using credibility
theory Monte Carlo simulation model, SpringerPlus 5 (1) (2016) 778.
[195] X. Cao, C. Gu, E. Zhao, Uncertainty instability risk analysis of high concrete arch
dam abutments, Math. Probl. Eng. (2017).
[196] ABS, Risk Assessment Application for the Marine and Offshore Oil and Gas
Industries, Technical Report. American Bureau of Shipping, Houston, USA, 2000.
[197] J.R. Harrald, I. Renda-Tanali, G.L. Shaw, C.B. Rubin, S. Yeletaysi, Review of risk
based prioritization/decision making methodologies for dams, Technical Report.
US army corps for engineers, The George Washington University, Institute for
Crisis, Disaster, and Risk Management, Washington, DC, 2004.
[198] D. Hartford, G. Baecher, Risk and uncertainty in dam safety, Technical Report.
Thomas Telford, Ltd, London, UK, 2004.
[199] R. Patev, C. Putcha, Development of fault trees for risk assessment of dam gates
and associated operating equipment, Int. J. Model. Simul. 25 (3) (2005) 190–201.
[200] D. Bowles, D. McClelland, Event tree analysis, Tech. Rep., Working paper prepared
for the Canadian Electricity Association Dam Safety Interest Group. Institute for
Dam Safety Risk Management, Utah State University, Logan, Utah, 2000.
[201] P. Hill, D. Bowles, R. Nathan, R. Herweynen, On the art of event tree modeling for
portfolio risk analyses, in: NZSOLD/ANCOLD 2001 Conference on Dams, 2001.
[202] A. Srivastava, Generalized event tree algorithm and software for dam safety risk
analysis, Technical Report. All Graduate Theses and Dissertations, Paper 32, USU,
Utah State University, Utah, 2008.
[203] Z. Micovic, D.N. Hartford, M.G. Schaefer, B.L. Barker, A non-traditional approach
to the analysis of flood hazard for dams, Stoch. Environ. Res. risk Assess. 30 (2)
(2016) 559–581.
[204] D. Hartford, G. Baecher, P. Zielinski, R. Patev, Operational Safety of Dams and
Reservoirs - Understanding the Reliability of Flow-Control Systems, ICE
Publishing, London, 2016.
[205] I. Gabriel-Martin, A. Sordo-Ward, L. Garrote, L.G. Castillo, Influence of initial
reservoir level and gate failure in dam safety analysis, Stoch. Approach, J. Hydrol.
550 (2017) 669–684.
[206] M.A. Hariri-Ardebili, Performance based earthquake engineering for concrete
dams (Ph.D. Thesis), University of Colorado, Boulder, CO, 2015.
[207] H. Tosun, İ. Zorluer, A. Orhan, E. Seyrek, H. Savaş, M. Türköz, Seismic hazard and
total risk analyses for large dams in Euphrates basin, Turk. Eng. Geol. 89 (1–2)
(2007) 155–170.
[208] A.S. Al-Homoud, A.I. Husein, Probabilistic assessment of seismic hazard of dam
sites in Jordan, Nat. Hazards 11 (2) (1995) 123–134.
[209] U.R. Vetter, J.P. Ake, R.C. Laforge, Seismic hazard evaluation for dams in northern
Colorado, USA, Nat. Hazards 14 (2–3) (1996) 227–240.
[210] I. Wong, P. Thomas, Comparing Seismic hazard estimates for dams in the US and
their implications to risk, in: Proceedings of the Association of Dam Safety Officials
Dam Safety. 2017.
[211] S. Pailoplee, Earthquake hazard of dams along the Mekong mainstream, Nat.
Hazards 74 (3) (2014) 1813–1827.
[212] E. Seyrek, H. Tosun, Deterministic approach to the seismic hazard of dam sites in
Kızılırmak basin, Turk. Nat. Hazards 59 (2) (2011) 787.
[213] R.E. Fat-Helbary, A.A. Tealb, A study of seismicity and earthquake hazard at the
proposed Kalabsha dam site, Aswan, Egypt, Nat. Hazards 25 (2) (2002) 117–133.
[214] J. García-Mayordomo, J. Insua-Arévalo, Seismic hazard assessment for the Itoiz
dam site (Western Pyrenees, Spain), Soil Dyn. Earthq. Eng. 31 (7) (2011)
1051–1063.
[215] E. Zacchei, J.L. Molina, R.M. Brasil, Seismic hazard and structural analysis of the
concrete arch dam (rules dam on Guadalfeo river), Procedia Eng. 199 (2017)
1332–1337.
[216] J.N. Brune, The seismic hazard at Tehri dam, Tectonophysics 218 (1–3) (1993)
281–286.
[217] R. Alipoor, M. Poorkermani, M. Zare, R. El Hamdouni, Active tectonic assessment
829
International Journal of Disaster Risk Reduction 31 (2018) 806–831
around Rudbar Lorestan dam site, High Zagros Belt (SW of Iran), Geomorphology
128 (1–2) (2011) 1–14.
[218] P. Zimmaro, J.P. Stewart, Probabilistic Seismic Hazard Analysis for a Dam Site in
Calabria (Southern Italy), Tech. Rep., University of California, Los Angeles,
Department of Civil and Environmental Engineering, Los Angeles, California,
2015.
[219] G. Fiorentino, L. Furgani, C. Nuti, F. Sabetta, Seismic Hazard and use of strong
motion time histories for dam seismic analyses, in: Proceedings of the 2nd
European Conference on Earthquake Engineering and Seismology, 2012.
[220] K. Pavlou, G. Kaviris, K. Chousianitis, G. Drakatos, V. Kouskouna, K. Makropoulos,
Seismic hazard assessment in Polyphyto Dam area (NW Greece) and its relation
with the unexpected earthquake of 13 May 1995 (Ms = 6.5, NW Greece), Nat.
Hazards Earth Syst. Sci. 13 (2013) 141–149.
[221] C. Johnson, D. Murphy, F. Makdisi, Seismic hazard analysis for qaraoun dam in the
bekaa valley, Lebanon, in: Proceedings of the Tenth U.S. National Conference on
Earthquake Engineering, Frontiers of Earthquake Engineering, 2014.
[222] M. Wieland, What seismic hazard information the dam engineers need from
seismologists and geologists?, in: Proceedings of the 2nd European Conference on
Earthquake Engineering and Seismology, 2014a.
[223] M. Wieland, Earthquake safety of existing dams, in: Proceedings of the First
European Conference on Earthquake Engineering and Seismology, Geneva,
Switzerland, 2006.
[224] M. Wieland, Seismic aspects of large dams, Keynote lecture, in: ANCOLD con-
ference, Adelaide, Australia, 2009.
[225] M. Wieland, Features of seismic hazard in large dam projects and strong motion
monitoring of large dams, Front. Archit. Civil. Eng. China 4 (1) (2010) 56–64.
[226] M. Wieland, ICOLD’s revised seismic design and performance criteria for large
storage dams, in: AFRICA 2013 - Water Storage and Hydropower Development for
Africa, Addis Ababa, Ethiopia, 2013.
[227] M. Wieland, Seismic hazard and seismic design and safety aspects of large dam
projects, in: Perspectives on European Earthquake Engineering and Seismology,
Springer, 627–650, 2014b.
[228] B. Frigo, A.P. Fantilli, V. De Biagi, G. Veylon, A protocol to assess the seismic
criticality of existing small concrete dams, Struct. Infrastruct. Eng. (2017) 1–10.
[229] R.E. Swain, D. Bowles, D. Ostenaa, A framework for characterization of extreme
floods for dam safety risk assessments, in: Proceedings of the 1998 USCOLD
Annual Lecture, 1998.
[230] USBR, A framework for characterizing extreme floods for dam safety risk assess-
ment, Technical Report. Utah State University and United States Department of
the Interior Bureau of Reclamation, Los Angeles, California, 1999.
[231] B. Ellingwood, P. Tekie, Fragility analysis of concrete gravity dams, ASCE J.
Infrastruct. Syst. 7 (2001) 41–48.
[232] R. Swain, J. England, K. Bullard, D. Raff, Hydrologic Hazard Curve Estimating
Procedures, Technical Report. U.S. Department of the Interior Bureau of
Reclamation, 2004.
[233] F. Lempérière, Dams and Floods, Engineering 3 (1) (2017) 144–149.
[234] N. Uddin, A.H. Ang, Quantitative Risk Assessment (QRA) for Natural Hazards,
American Society of Civil Engineers, 2011.
[235] FEMA, Federal guidelines for dam safety risk management, Technical Report.
Federal Emergency Management Agency, Washington, DC, 2015.
[236] P. Gardoni, C. Murphy, A scale of risk, Risk Anal. 34 (7) (2014) 1208–1227.
[237] D.S. Bowles, What is ALARP and how can it improve dam safety decisions?, in:
ASDSO conference On Dams, 2013a.
[238] HSE, Reducing risks, protecting people: HSE’s Decisionmaking process, Risk as-
sessment policy unit, Technical Report. Health and Safety Executive (HSE),
London, England, 2001.
[239] D.F. Munger, D.S. Bowles, D.D. Boyer, D.W. Davis, D.A. Margo, D.A. Moser, P.J.
Regan, N. Snorteland, Interim tolerable risk guidelines for US Army Corps of
Engineers dams, in: Proceedings of the 29 th Annual United States Society on
Dams Conference on Managing our Water Retention Systems:, 2009. pp.
1125–1141.
[240] W. Ge, Z. Li, R.Y. Liang, W. Li, Y. Cai, Methodology for establishing risk criteria for
dams in developing countries, case study of China, Water Resour. Manag. 31 (13)
(2017) 4063–4074.
[241] S.-T. Cheng, B.C. Yen, W.H. Tang, Overtopping risk for an existing dam, Tech. Rep.
UILU-ENG-82-2017, Departement of Civil Engineering, University of Illinois at
Urbana-Champaign, Urbana, Illinois, 1982.
[242] National Research Council, Estimating probabilities of extreme floods: methods
and recommended research, National Academies, 1988.
[243] D. Bowles, Verde River risk assessment: An interim solution study, in: Invited
lecture in the proceedings of the U.S. Committee on Large Dams Annual Meeting,
Phoenix, Arizona, 1988.
[244] J. Stedinger, D.C. Heath, K. Thompson, Risk Analysis for Dam Safety Evaluation:
Hydrologic Risk., Technical Report. Army engineer inst for water resources Fort
Belvoir VA, 1996.
[245] D.S. Bowles, L.R. Anderson, T.F. Glover, A role for risk assessment in dam safety
management, in: Proceedings of the 3rd International Conference on Hydropower,
vol. 97, 1997.
[246] M. Kostov, H. Boncheva, D. Stefanov, G. Varbanov, A. Kaneva, N. Koleva, Seismic
risk assessment of large concrete gravity dams, in: Proceedings of the 11th
European Conference on Earthquake Engineering, Paris, 1998.
[247] M. Kostov, G. Varbanov, A. Kaneva, N. Koleva, Seismic risk of large dams in
Bulgaria, in: Proceedings of the 14th World Conference on Earthquake
Engineering, 2008.
[248] D.S. Bowles, L.R. Anderson, T.F. Glover, S.S. Chauhan, Portfolio risk assessment: A
tool for dam safety risk management, in: Proceedings of USCOLD 1998 Annual
M.A. Hariri-Ardebili
Lecture. Buffalo, New York.
[249] D. Bowles, Advances in the practice and use of portfolio risk assessment, in:
Proceedings of the ANCOLD 2000 Conference on Dams, 21-32, 2001.
[250] D. Bowles, Overview of portfolio dam safety risk management, in: Proceedings of
the 81st Annual Meeting of the International Commission on Large Dams. Seattle.
United States.[Links], 2013b.
[251] D.S. Bowles, L.R. Anderson, J.B. Evelyn, T.F. Glover, D.M. Van Dorpe, Alamo dam
demonstration risk assessment, in: Proceedings of the Australian Committee on
Large Dams (ANCOLD) Annual Meeting, Jindabyne, Australia, 1999. pp.113–128.
[252] K. Rettermeier, B. Nilkens, B. Falkenhagen, J. Kongeter, New developments in dam
safety-feasibility evaluation on risk assessment, institute of hydraulic engineering
and water resources management, Aachen University of Technology, Aachen,
Germany.
[253] K.S. Richards, N.J. Snorteland, Risk management: challenges and practice for US
Dam and levee safety, Geoenvironmental Practices and Sustainability, Springer,
2017, pp. 173–179.
[254] P. Hill, D. Bowles, P. Jordan, R. Nathan, Estimating overall risk of dam failure:
practical considerations in combining failure probabilities, in: Proceedings of the
ANCOLD Risk Workshop, 2003.
[255] S.S. Chauhan, D.S. Bowles, Dam safety risk assessment with uncertainty analysis,
in: Proceedings of the Australian Committee on Large Dams Risk Workshop,
Launceston, Australia, 2003.
[256] L. Peyras, P. Royet, D. Boissier, Dam ageing diagnosis and risk analysis: devel-
opment of methods to support expert judgment, Can. Geotech. J. 43 (2) (2006)
169–186.
[257] R.V. Matalucci, Risk Assessment Metohodology for Dams (RAM-D SM), in:
Proceedings of the 6th International Conference on Probabilistic Safety
Assessment and Management, San Juan, Puerto Rico, USA, 2002. pp.69–176.
[258] M. Meghella, M. Eusebio, A risk assessment tool to effectively support decision
makers to prioritize maintenance, repair and upgrading of dams, in: Dam
Maintenance and Rehabilitation, 159-168, 2003.
[259] D.S. Bowles, M. Aboelata, Evacuation and life-loss estimation model for natural
and dam break floods, Extreme hydrological Events: New Concepts for Security,
Springer, 2006, pp. 363–383.
[260] X. He, C. Gu, Z. Wu, H. Su, Dam risk assistant analysis system design, Sci. China
Ser. E Technol. Sci. 51 (2) (2008) 101–109.
[261] A. Srivastava, A Computational Framework for Dam Safety Risk Assessment With
Uncertainty Analysis (Ph.D. thesis), Utah State University, 2013.
[262] A. Srivastava, D.S. Bowles, S.S. Chauhan, DAMRAE-U: A tool for including un-
certainty in dam safety risk assessment, in: Proceedings of the ASDSO Conference
on Dams, Denver, USA, 2012.
[263] iPresas, iPresas Risk Analysis, 〈http://ipresas.com/en/〉, last viewed July 2018,
2014.
[264] A. Serrano-Lombillo, A. Morales-Torres, I. Escuder-Bueno, L. Altarejos-García, A
new risk reduction indicator for dam safety management combining efficiency and
equity principles, Struct. Infrastruct. Eng. 13 (9) (2017) 1157–1166.
[265] M. Smith, Dam risk analysis using Bayesian networks, in: Proceedings of the
Geohazards, ECI Symposium Series, 2006.
[266] Z. Wu, H. Su, H. Guo, Assessment model of dam operation risk based on mon-
itoring data, Sci. China Ser. E Technol. Sci. 50 (1) (2007) 144–152.
[267] Z. Wu, H. Su, H. Guo, Risk assessment method of major unsafe hydroelectric
project, Sci. China Ser. E: Technol. Sci. 51 (9) (2008) 1345–1352.
[268] A. Serrano-Lombillo, I. Escuder-Bueno, M.G. de Membrillera-Ortuño, L. Altarejos-
Garcia, Methodology for the calculation of annualized incremental risks in systems
of dams, Risk Anal. 31 (6) (2011) 1000–1015.
[269] H. Su, Z. Wen, Interval risk analysis for gravity dam instability, Eng. Fail. Anal. 33
(2013) 83–96.
[270] H. Su, Z. Wen, J. Hu, Z. Wu, Evaluation model for service life of dam based on
time-varying risk probability, Sci. China Ser. E Technol. Sci. 52 (7) (2009)
1966–1973.
[271] G. Cloete, J. Retief, C. Viljoen, A rational quantitative optimal approach to dam
safety risk reduction, Civil. Eng. Environ. Syst. 33 (2) (2016) 85–105.
[272] E. Brown, Reducing risks in the investigation, design and construction of large
concrete dams, J. Rock. Mech. Geotech. Eng. 9 (2) (2017) 197–209.
[273] J.T. Castillo-Rodríguez, J.T. Needham, A. Morales-Torres, I. Escuder-Bueno, A
combined risk analysis approach for complex dam-levee systems, Struct.
Infrastruct. Eng. 13 (12) (2017) 1624–1638.
[274] K. Porter, An overview of PEER’s performance-based earthquake engineering
methodology, in: Proceedings of the 9th International Conference on Applications
of Statistics and Probability in Civil Engineering (ICASP9), San Francisco, CA,
2003.
[275] B. Ellingwood, K. Kinali, Quantifying and communicating uncertainty in seismic
risk assessment, Struct. Saf. 31 (2) (2009) 179–187 (risk Acceptance and Risk
CommunicationRisk Acceptance and Risk Communication).
[276] F. Jalayer, P. Franchin, P. Pinto, A scalar damage measure for seismic reliability
analysis of RC frames, Earthq. Eng. Struct. Dyn. 36 (2007) 2059–2079.
[277] E. Bojorquez, I. Iervolino, A. Reyes-Salazar, S. Ruiz, Comparing vector-valued
intensity measures for fragility analysis of steel frames in the case of narrow-band
ground motions, Eng. Struct. 45 (2012) 472–480.
[278] M.A. Hariri-Ardebili, V.E. Saouma, Seismic fragility analysis of concrete dams: a
state-of-the-art review, Eng. Struct. 128 (2016) 374–399.
[279] J. de Araújo, A. Awruch, Probabilistic finite element analysis of concrete gravity
dams, Adv. Eng. Softw. 29 (1998) 97–104.
[280] K. Porter, R. Kennedy, R. Bachman, Creating fragility functions for performance-
based earthquake engineering, Earthq. Spectra 23 (2007) 471–489.
[281] J. Baker, Efficient analytical fragility function fitting using dynamic structural
830
International Journal of Disaster Risk Reduction 31 (2018) 806–831
analysis, Earthq. Spectra 31 (1) (2015) 579–599.
[282] P. Tekie, B. Ellingwood, Seismic fragility assessment of concrete gravity dams,
Earthq. Eng. Struct. Dyn. 32 (2003) 2221–2240.
[283] P. Tekie, B. Ellingwood, Fragility analysis of concrete gravity dams, Tech. Rep.,
Georgia Institute of Technology, Prepared for U.S. Army Corps of Engineers,
Washington, DC, 2002.
[284] L. Lin, J. Adams, Lessons for the fragility of Canadian hydropower components
under seismic loading, in: Proceedings of the 9th Canadian Conference on
Earthquake Engineering, Ottawa, Ontario, Canada, 2007. pp. 1762–1771.
[285] L. Lin, J. Adams, Seismic vulnerability and prioritization ranking of dams in
Canada, in: Proceedings of the 14th World Conference on Earthquake Engineering,
Beijing, China, 2008.
[286] ATC-13, Earthquake damage evaluation data for California. Report No.ATC-13,
Technical Report. Applied Technology Council, Redwood City, CA, 1985.
[287] S. Mirzahosseinkashani, M. Ghaemian, Seismic fragility assessment of concrete
gravity dams, in: Proceedings of the 29th Annual USSD Conference, Nashville,
Tennessee, US, 2009.
[288] A. Lupoi, C. Callari, The role of probabilistic methods in evaluating the seismic risk
of concrete dams, in: M. Dolšek (Ed.), Protection of Built Environment Against
Earthquakes, Springer, 2011, pp. 309–329.
[289] A. Lupoi, C. Callari, A probabilistic method for the seismic assessment of existing
concrete gravity dams, Struct. Infrastruct. Eng. 8 (2012) 985–998.
[290] X. Yao, A. Elnashai, J. Jiang, Analytical seismic fragility analysis of concrete arch
dams, in: Proceedings of the 15th World Conference on Earthquake Engineering,
Lisbon, Portugal, 2012.
[291] H. Abdelhamid, B. Mahmoud, M. Hussein, Seismic fragility and uncertainty ana-
lysis of concrete gravity dams under near-fault ground motions, Civil. Environ.
Res. 5 (2013) 123–129.
[292] B. Ju, W. Jung, Evaluation of seismic fragility of weir structures in South Korea,
Math. Probl. Eng. 2015 (2015) 10.
[293] Y. Ghanaat, P. Hashimoto, O. Zuchuat, R. Kennedy, Seismic fragility of mühleberg
dam using nonlinear analysis with latin hypercube simulation, in: Proceeding of
the 2011 USSD Annual Conference, 2011. pp.1197–1212.
[294] Y. Ghanaat, R. Patev, A. Chudgar, Seismic fragility analysis of concrete gravity
dams, in: Proceedings of the 15th World Conference on Earthquake Engineering,
Lisbon, Portugal, 2012.
[295] Y. Ghanaat, R. Patev, A. Chudgar, Seismic fragility for risk assessment of concrete
gravity dams, in: Proceeding of the 2015 USSD Annual Conference, 2015. pp.
645–660.
[296] V. Kadkhodayan, M. Aghajanzadeh, H. Mirzabozorg, Seismic assessment of arch
dams using fragility curves, Civil. Eng. J. 1 (2) (2015) 14–20.
[297] M.A. Hariri-Ardebili, V. Saouma, K.A. Porter, Quantification of seismic potential
failure modes in concrete dams, Earthq. Eng. Struct. Dyn. 45 (2016) 979–997.
[298] M.A. Hariri-Ardebili, V. Saouma, Probabilistic seismic demand model and optimal
intensity measure for concrete dams, Struct. Saf. 59 (2016) 67–85.
[299] A. Cornell, F. Jalayer, R. Hamburger, Probabilistic basis for SAC federal emergency
management agency steel moment frame guidelines, J. Struct. Eng. 128 (2002)
526–532.
[300] J. Padgett, B. Nielson, R. DesRoches, Selection of optimal intensity measures in
probabilistic seismic demand models of highway bridge portfolios, Earthq. Eng.
Struct. Dyn. 37 (2008) 711–725.
[301] M. Hariri-Ardebili, V. Saouma, Collapse fragility curves for concrete dams: com-
prehensive study, ASCE J. Struct. Eng. 142 (10) (2016) 04016075.
[302] D. Vamvatsikos, C. Cornell, Incremental dynamic analysis, Earthq. Eng. Struct.
Dyn. 31 (2002) 491–514.
[303] C. Bernier, J.E. Padgett, J. Proulx, P. Paultre, Seismic fragility of concrete gravity
dams with spatial variation of angle of friction: case study, J. Struct. Eng. (2015)
05015002.
[304] C. Bernier, R. Monteiro, P. Paultre, Using the conditional spectrum method for
improved fragility assessment of concrete gravity dams in Eastern Canada, Earthq.
Spectra 0 (0) (2016).
[305] M.I. Ansari, P. Agarwal, Categorization of damage index of concrete gravity dam
for the health monitoring after earthquake, J. Earthq. Eng. 0 (ja) (2016).
[306] M.I. Ansari, M. Saqib, P. Agarwal, Geometric Configuration Effects on Nonlinear
Seismic Behavior of Concrete Gravity Dam, J. Earthq. Tsunami (2018) 1850003.
[307] A. Morales-Torres, I. Escuder-Bueno, L. Altarejos-Garcia, A. Serrano-Lombillo,
Building fragility curves of sliding failure of concrete gravity dams integrating
natural and epistemic uncertainties, Eng. Struct. 125 (2016) 227–235.
[308] Y. Yazdani, M. Alembagheri, Seismic vulnerability of gravity dams in near-fault
areas, Soil Dyn. Earthq. Eng. 102 (2017) 15–24.
[309] R. Pang, B. Xu, X. Kong, D. Zou, Seismic fragility for high CFRDs based on de-
formation and damage index through incremental dynamic analysis, Soil Dyn.
Earthq. Eng. 104 (2018) 432–436.
[310] M. Alembagheri, Investigating efficiency of vector-valued intensity measures in
seismic demand assessment of concrete dams, Adv. Civil. Eng. (2018).
[311] M.A. Hariri-Ardebili, V.E. Saouma, Random Response Spectrum Analysis of
Gravity Dam Classes: Simplified, Practical and Fast Approach, Earthquake Spectra
34 (2) (2018) 941–975, https://doi.org/10.1193/021517EQS033M.
[312] J.-T. Wang, M.-X. Zhang, A.-Y. Jin, C.-H. Zhang, Seismic fragility of arch dams
based on damage analysis, Soil Dyn. Earthq. Eng. 109 (2018) 58–68.
[313] T.E. Downing, L. Ringius, M. Hulme, D. Waughray, Adapting to climate change in
Africa, Mitig. Adapt. Strateg. Glob. Change 2 (1) (1997) 19–44.
[314] Y. Lei, Y. Yue, H. Zhou, W. Yin, et al., Rethinking the relationships of vulnerability,
resilience, and adaptation from a disaster risk perspective, Nat. Hazards 70 (1)
(2014) 609–627.
[315] J.X. Kasperson, R.E. Kasperson, The social contours of risk: publics, risk
M.A. Hariri-Ardebili
communication and the social amplification of risk, vol. 1, Earthscan, 2005.
[316] M. Parry, O. Canziani, J. Palutikof, P.J. van der Linden, C.E. Hanson, et al.,
Climate change 2007: impacts, adaptation and vulnerability, vol. 4, Cambridge
University Press Cambridge, 2007.
[317] B.L. Turner, R.E. Kasperson, P.A. Matson, J.J. McCarthy, R.W. Corell,
L. Christensen, N. Eckley, J.X. Kasperson, A. Luers, M.L. Martello, et al., A fra-
mework for vulnerability analysis in sustainability science, Proc. Natl. Acad. Sci.
USA 100 (14) (2003) 8074–8079.
[318] S.L. Cutter, B.J. Boruff, W.L. Shirley, Social vulnerability to environmental ha-
zards, Social. Sci. Q. 84 (2) (2003) 242–261.
[319] P. Blaikie, T. Cannon, I. Davis, B. Wisner, At Risk: Natural Hazards, People's
Vulnerability and Disasters, Routledge, 2014.
[320] H. Zhou, J. Wan, H. Jia, et al., Resilience to natural hazards: a geographic per-
spective, Nat. Hazards 53 (1) (2010) 21–41.
[321] G. Han, Understanding Regional Dynamics of Vulnerability: A Historical Approach
to the Flood Problem in China, (Ph.D. thesis), Clark University, Graduate School of
Geography, 2011.
[322] K. Porter, A beginner’s guide to fragility, vulnerability, and risk, in: M. Beer,
I.A. Kougioumtzoglou, E. Patelli, I.S.-K. Au (Eds.), Encyclopedia of Earthquake
Engineering, Springer Berlin Heidelberg, 2015, pp. 1–29.
[323] A. Muntasir Billah, M. Shahria Alam, Seismic fragility assessment of highway
bridges: a state-of-the-art review, Struct. Infrastruct. Eng. 11 (6) (2015) 804–832.
[324] G.B. Baecher, M.E. Paté, R. De Neufville, Risk of dam failure in benefit-cost ana-
lysis, Water Resour. Res. 16 (3) (1980) 449–456.
[325] M.E. Paté-Cornell, G. Tagaras, Risk costs for new dams: economic analysis and
effects of monitoring, Water Resour. Res. 22 (1) (1986) 5–14.
[326] L.B. Lave, D. Resendiz-Carrillo, F.C. Mcmichael, Safety goals for high-hazard
dams: are dams too safe? Water Resour. Res. 26 (7) (1990) 1383–1391.
[327] B. Ellingwood, R.B. Corotis, J. Boland, N.P. Jones, Assessing cost of dam failure, J.
Water Resour. Plan. Manag. 119 (1) (1993) 64–82.
[328] A. Ansar, B. Flyvbjerg, A. Budzier, D. Lunn, Should we build more large dams? The
actual costs of hydropower megaproject development, Energy Policy 69 (2014)
43–56.
[329] A. Morales-Torres, A. Serrano-Lombillo, I. Escuder-Bueno, L. Altarejos-García, The
suitability of risk reduction indicators to inform dam safety management, Struct.
Infrastruct. Eng. 12 (11) (2016) 1465–1476.
[330] J. Pisaniello, J. Tingey-Holyoak, Best practice land use planning and cost-effective
tools to assure safety downstream of private dams, Water Resour. 43 (4) (2016)
730–742.
[331] Presidential Policy Directive, Presidential Policy Directive – Critical Infrastructure
Security and Resilience, 〈https://obamawhitehouse.archives.gov/the-press-
office/2013/02/12/presidential-policy-directive-critical-infrastructure-security-
and-resil〉, last viewed January 2018, 2013.
[332] M. Bruneau, A. Reinhorn, Exploring the concept of seismic resilience for acute care
facilities, Earthq. Spectra 23 (1) (2007) 41–62.
[333] G.P. Cimellaro, A.M. Reinhorn, M. Bruneau, Framework for analytical quantifi-
cation of disaster resilience, Eng. Struct. 32 (11) (2010) 3639–3649.
[334] P. Bocchini, D.M. Frangopol, T. Ummenhofer, T. Zinke, Resilience and sustain-
ability of civil infrastructure: toward a unified approach, J. Infrastruct. Syst. 20 (2)
(2013) 04014004.
[335] G. Deodatis, B.R. Ellingwood, D.M. Frangopol, Safety, Reliability, Risk and Life-
cycle Performance of Structures and Infrastructures, CRC Press, 2014.
[336] N. Sharma, A. Tabandeh, P. Gardoni, Resilience analysis: a mathematical for-
mulation to model resilience of engineering systems, Sustain. Resilient Infrastruct.
(2017) 1–19.
[337] B.M. Ayyub, Resilience metrics for multi-hazard environments, Vulnerability,
Uncertainty, and Risk: Quantification, Mitigation, and Management, American
Society of Civil Engineers, 2014, pp. 925–936.
[338] F. Hossain, J. Arnold, E. Beighley, C. Brown, S. Burian, J. Chen, S. Madadgar,
A. Mitra, D. Niyogi, R. Pielke Sret al., Local-to-regional landscape drivers of ex-
treme weather and climate: implications for water infrastructure resilience, J.
Hydrol. Eng. 20 (7) (2015) 02515002.
[339] M. Bruneau, S.E. Chang, R.T. Eguchi, G.C. Lee, T.D. O'Rourke, A.M. Reinhorn,
M. Shinozuka, K. Tierney, W.A. Wallace, D. Von Winterfeldt, A framework to
quantitatively assess and enhance the seismic resilience of communities, Earthq.
Spectra 19 (4) (2003) 733–752.
[340] S.L. Cutter, J.A. Ahearn, B. Amadei, P. Crawford, E.A. Eide, G.E. Galloway,
M.F. Goodchild, H.C. Kunreuther, M. Li-Vollmer, M. Schoch-Spana, et al., Disaster
resilience: a national imperative, Environ.: Sci. Policy Sustain. Dev. 55 (2) (2013)
25–29.
831
International Journal of Disaster Risk Reduction 31 (2018) 806–831
[341] G. Cimellaro, A. Reinhorn, M. Bruneau, Quantification of seismic resilience,
in: Proceedings of the 8th US National Conference on Earthquake Engineering,
2006. pp.18–22.
[342] H. Bonstrom, R.B. Corotis, First-order reliability approach to quantify and improve
building portfolio resilience, J. Struct. Eng. 142 (8) (2016) C4014001.
[343] Y.Y. Haimes, On the definition of resilience in systems, Risk Anal. 29 (4) (2009)
498–501.
[344] S.B. Manyena, S. Gordon, Bridging the concepts of resilience, fragility and stabi-
lisation, Disaster Prev. Manag. 24 (1) (2015) 38–52.
[345] R.B. Corotis, Conceptual and analytical differences between resiliency and relia-
bility for seismic hazards, in: Proceedings of the Structures Congress, 2010–2020,
2011.
[346] W.J. Graham, Should dams be modified for the probable maximum flood?, JAWR,
A J. Am. Water Resour. Assoc. 36 (5) (2000) 953–963.
[347] F. Hossain, E. Beighley, S. Burian, J. Chen, A. Mitra, D. Niyogi, R. Pielke Sr,
D. Wegner, Review of approaches and recommendations for improving resilience
of water management infrastructure: the case for large dams, J. Infrastruct. Syst.
23 (4) (2017) 02517001.
[348] National Research Council, Dam and Levee Safety and Community Resilience: A
Vision for Future Practice, National Academies Press, 2012.
[349] B. Kim, S.C. Shin, D.Y. Kim, Resilience assessment of dams' flood-control service, J.
Korean Soc. Civil. Eng. 34 (6) (2014) 1919–1924.
[350] M. Karamouz, A. Zeynolabedin, M. Olyaei, Regional drought resiliency and vul-
nerability, J. Hydrol. Eng. 21 (11) (2016) 05016028.
[351] C. Lee, J. Jung, J. An, J.Y. Kim, Y. Choi, Establishment of resilient infrastructures
for the mitigation of an urban water problem: 1. Robustness assessment of struc-
tural alternatives for the problem of urban floods, Ecol. Resilient Infrastruct. 3 (2)
(2016) 117–125.
[352] T.V. Bheemasetti, A.J. Puppala, L. Verreault, A. Pedarla, Y.P. Weatherton,
Optimising geotechnical data for analysis of levee resilience, in: Proceedings of the
Institution of Civil Engineers-Engineering Sustainability, Thomas Telford Ltd,
2017. pp.1–12.
[353] B. Kim, S.C. Shin, et al., A resilience loss assessment framework for evaluating
flood-control dam safety upgrades, Nat. Hazards 86 (2) (2017) 805–819.
[354] N. Matthews, M. McCartney, Opportunities for building resilience and lessons for
navigating risks: dams and the water energy food nexus, Environ. Prog. Sustain.
Energy 37 (1) (2018) 56–61.
[355] G.H. Brundtland, Our common future Call for action, Environ. Conserv. 14 (4)
(1987) 291–294.
[356] K.R. Mackie, M. Kucukvar, O. Tatari, A. Elgamal, Sustainability metrics for per-
formance-based seismic bridge response, J. Struct. Eng. 142 (8) (2015) C4015001.
[357] Y. Li, Q. He, J. Han, X. Luo, Resilience and sustainability: an integrated method for
quantitative assessment of a bridge-a case study, WIT Trans. Built Environ. 164
(2016) 25–37.
[358] D. Marchese, E. Reynolds, M.E. Bates, H. Morgan, S.S. Clark, I. Linkov, Resilience
and sustainability: similarities and differences in environmental management
applications, Sci. Total Environ. 613 (2018) 1275–1283.
[359] D. Basu, A. Misra, A.J. Puppala, Sustainability and geotechnical engineering:
perspectives and review, Can. Geotech. J. 52 (1) (2014) 96–113.
[360] J.T. Das, A.J. Puppala, T.V. Bheemasetti, L.A. Walshire, M.K. Corcoran,
Sustainability and resilience analyses in slope stabilisation, in: Proceedings of the
Institution of Civil Engineers-Engineering Sustainability, Thomas Telford Ltd,
2016. pp.1–12.
[361] S.P. Sevionovic, Risk in sustainable water resources management, in: Proceedings
of Rabat symposium sustainable water resources under increasing uncertainty.
IAHS Publication, 3-17, 1997.
[362] I. Yüksel, Dams and hydropower for sustainable development, Energy Sources,
Part B 4 (1) (2009) 100–110.
[363] M. Yang, X. Qian, Y. Zhang, J. Sheng, D. Shen, Assessing alternatives for sus-
tainable management of a flood control dam, Procedia Environ. Sci. 2 (2010)
98–110.
[364] E. Bird, The Socioeconomic Impact of Hydroelectric Dams on Developing
Communities: A Case Study of the Chalillo Dam and the Communities of the Macal
River Valley, Cayo District, Belize, Central America (Master’s thesis), University of
Vermont, USA, 2012.
[365] Z. Tessler, C.J. Vörösmarty, M. Grossberg, I. Gladkova, H. Aizenman, J. Syvitski,
E. Foufoula-Georgiou, Profiling risk and sustainability in coastal deltas of the
world, Science 349 (6248) (2015) 638–643.
[366] M. Wieland, Safety aspects of sustainable storage dams and earthquake safety of
existing dams, Engineering 2 (3) (2016) 325–331.